diff --git a/include/net/net_stats.h b/include/net/net_stats.h
index 89b0d05f1f1..8b524d7f24e 100644
--- a/include/net/net_stats.h
+++ b/include/net/net_stats.h
@@ -22,6 +22,11 @@ extern "C" {
 
 typedef u32_t net_stats_t;
 
+struct net_stats_bytes {
+	u32_t sent;
+	u32_t received;
+};
+
 struct net_stats_ip {
 	/** Number of received packets at the IP layer. */
 	net_stats_t recv;
@@ -78,6 +83,12 @@ struct net_stats_icmp {
 };
 
 struct net_stats_tcp {
+	/** Amount of received and sent TCP application data. */
+	struct net_stats_bytes bytes;
+
+	/** Amount of retransmitted data. */
+	net_stats_t resent;
+
 	/** Number of recived TCP segments. */
 	net_stats_t recv;
 
@@ -90,22 +101,25 @@ struct net_stats_tcp {
 	/** Number of TCP segments with a bad checksum. */
 	net_stats_t chkerr;
 
-	/** Number of TCP segments with a bad ACK number. */
+	/** Number of received TCP segments with a bad ACK number. */
 	net_stats_t ackerr;
 
+	/** Number of received bad TCP RST (reset) segments. */
+	net_stats_t rsterr;
+
 	/** Number of received TCP RST (reset) segments. */
 	net_stats_t rst;
 
 	/** Number of retransmitted TCP segments. */
 	net_stats_t rexmit;
 
-	/** Number of dropped SYNs because too few connections were
-	 * available.
+	/** Number of dropped connection attempts because too few connections
+	 * were available.
 	 */
-	net_stats_t syndrop;
+	net_stats_t conndrop;
 
-	/** Number of SYNs for closed ports, triggering a RST. */
-	net_stats_t synrst;
+	/** Number of connection attempts for closed ports, triggering a RST. */
+	net_stats_t connrst;
 };
 
 struct net_stats_udp {
@@ -207,11 +221,6 @@ struct net_stats_ipv6_mld {
 	net_stats_t drop;
 };
 
-struct net_stats_bytes {
-	u32_t sent;
-	u32_t received;
-};
-
 struct net_stats {
 	net_stats_t processing_error;
 
diff --git a/subsys/net/ip/Kconfig.stats b/subsys/net/ip/Kconfig.stats
index 40ee3a89ce0..0bada1c7e72 100644
--- a/subsys/net/ip/Kconfig.stats
+++ b/subsys/net/ip/Kconfig.stats
@@ -69,7 +69,7 @@ config NET_STATISTICS_UDP
 config NET_STATISTICS_TCP
 	bool "TCP statistics"
 	depends on NET_TCP
-	default n
+	default y
 	help
 	Keep track of TCP related statistics
 
diff --git a/subsys/net/ip/connection.c b/subsys/net/ip/connection.c
index 35db99e7307..4d8968db7dd 100644
--- a/subsys/net/ip/connection.c
+++ b/subsys/net/ip/connection.c
@@ -843,9 +843,7 @@ enum net_verdict net_conn_input(enum net_ip_protocol proto, struct net_pkt *pkt)
 			goto drop;
 		}
 
-		if (proto == IPPROTO_UDP) {
-			net_stats_update_udp_recv();
-		}
+		net_stats_update_per_proto_recv(proto);
 
 		return NET_OK;
 	}
@@ -871,12 +869,14 @@ enum net_verdict net_conn_input(enum net_ip_protocol proto, struct net_pkt *pkt)
 #endif
 	{
 		send_icmp_error(pkt);
+
+		if (IS_ENABLED(CONFIG_NET_TCP) && proto == IPPROTO_TCP) {
+			net_stats_update_tcp_seg_connrst();
+		}
 	}
 
 drop:
-	if (proto == IPPROTO_UDP) {
-		net_stats_update_udp_drop();
-	}
+	net_stats_update_per_proto_drop(proto);
 
 	return NET_DROP;
 }
diff --git a/subsys/net/ip/net_context.c b/subsys/net/ip/net_context.c
index 2f8a2952db0..f95230b1b2f 100644
--- a/subsys/net/ip/net_context.c
+++ b/subsys/net/ip/net_context.c
@@ -32,6 +32,7 @@
 #include "ipv4.h"
 #include "udp.h"
 #include "tcp.h"
+#include "net_stats.h"
 
 #define NET_MAX_CONTEXT CONFIG_NET_MAX_CONTEXTS
 
@@ -823,9 +824,12 @@ NET_CONN_CB(tcp_established)
 	if (tcp_flags & NET_TCP_RST) {
 		/* We only accept RST packet that has valid seq field. */
 		if (!net_tcp_validate_seq(context->tcp, pkt)) {
+			net_stats_update_tcp_seg_rsterr();
 			return NET_DROP;
 		}
 
+		net_stats_update_tcp_seg_rst();
+
 		net_tcp_print_recv_info("RST", pkt, NET_TCP_HDR(pkt)->src_port);
 
 		if (context->recv_cb) {
@@ -917,9 +921,12 @@ NET_CONN_CB(tcp_synack_received)
 	if (NET_TCP_FLAGS(pkt) & NET_TCP_RST) {
 		/* We only accept RST packet that has valid seq field. */
 		if (!net_tcp_validate_seq(context->tcp, pkt)) {
+			net_stats_update_tcp_seg_rsterr();
 			return NET_DROP;
 		}
 
+		net_stats_update_tcp_seg_rst();
+
 		if (context->connect_cb) {
 			context->connect_cb(context, -ECONNREFUSED,
 					    context->user_data);
@@ -1350,9 +1357,12 @@ NET_CONN_CB(tcp_syn_rcvd)
 
 		/* We only accept RST packet that has valid seq field. */
 		if (!net_tcp_validate_seq(tcp, pkt)) {
+			net_stats_update_tcp_seg_rsterr();
 			return NET_DROP;
 		}
 
+		net_stats_update_tcp_seg_rst();
+
 		ack_timer_cancel(tcp);
 
 		net_tcp_print_recv_info("RST", pkt, NET_TCP_HDR(pkt)->src_port);
@@ -1399,7 +1409,7 @@ NET_CONN_CB(tcp_syn_rcvd)
 		if (ret < 0) {
 			NET_DBG("Cannot get accepted context, "
 				"connection reset");
-			goto reset;
+			goto conndrop;
 		}
 
 		new_context->tcp->recv_max_ack = context->tcp->recv_max_ack;
@@ -1460,7 +1470,7 @@ NET_CONN_CB(tcp_syn_rcvd)
 			NET_DBG("Cannot bind accepted context, "
 				"connection reset");
 			net_context_unref(new_context);
-			goto reset;
+			goto conndrop;
 		}
 
 		new_context->flags |= NET_CONTEXT_REMOTE_ADDR_SET;
@@ -1479,7 +1489,7 @@ NET_CONN_CB(tcp_syn_rcvd)
 			NET_DBG("Cannot register accepted TCP handler (%d)",
 				ret);
 			net_context_unref(new_context);
-			goto reset;
+			goto conndrop;
 		}
 
 		/* Swap the newly-created TCP states with the one that
@@ -1512,6 +1522,9 @@ NET_CONN_CB(tcp_syn_rcvd)
 
 	return NET_DROP;
 
+conndrop:
+	net_stats_update_tcp_seg_conndrop();
+
 reset:
 	{
 		struct sockaddr peer;
@@ -1930,6 +1943,8 @@ static enum net_verdict packet_received(struct net_conn *conn,
 		net_pkt_appdata(pkt), net_pkt_appdatalen(pkt),
 		net_pkt_get_len(pkt));
 
+	net_stats_update_tcp_recv(net_pkt_appdatalen(pkt));
+
 	context->recv_cb(context, pkt, 0, user_data);
 
 #if defined(CONFIG_NET_CONTEXT_SYNC_RECV)
diff --git a/subsys/net/ip/net_if.c b/subsys/net/ip/net_if.c
index 07c5ab8624e..6d06b1b5455 100644
--- a/subsys/net/ip/net_if.c
+++ b/subsys/net/ip/net_if.c
@@ -70,8 +70,15 @@ static inline void net_context_send_cb(struct net_context *context,
 #if defined(CONFIG_NET_UDP)
 	if (net_context_get_ip_proto(context) == IPPROTO_UDP) {
 		net_stats_update_udp_sent();
-	}
+	} else
 #endif
+#if defined(CONFIG_NET_TCP)
+	if (net_context_get_ip_proto(context) == IPPROTO_TCP) {
+		net_stats_update_tcp_seg_sent();
+	} else
+#endif
+	{
+	}
 }
 
 static bool net_if_tx(struct net_if *iface)
diff --git a/subsys/net/ip/net_shell.c b/subsys/net/ip/net_shell.c
index ecf38aa3036..014f35be532 100644
--- a/subsys/net/ip/net_shell.c
+++ b/subsys/net/ip/net_shell.c
@@ -373,6 +373,27 @@ static inline void net_shell_print_statistics(void)
 	       GET_STAT(udp.chkerr));
 #endif
 
+#if defined(CONFIG_NET_STATISTICS_TCP)
+	printk("TCP bytes recv %u\tsent\t%d\n",
+	       GET_STAT(tcp.bytes.received),
+	       GET_STAT(tcp.bytes.sent));
+	printk("TCP seg recv   %d\tsent\t%d\tdrop\t%d\n",
+	       GET_STAT(tcp.recv),
+	       GET_STAT(tcp.sent),
+	       GET_STAT(tcp.drop));
+	printk("TCP seg resent %d\tchkerr\t%d\tackerr\t%d\n",
+	       GET_STAT(tcp.resent),
+	       GET_STAT(tcp.chkerr),
+	       GET_STAT(tcp.ackerr));
+	printk("TCP seg rsterr %d\trst\t%d\tre-xmit\t%d\n",
+	       GET_STAT(tcp.rsterr),
+	       GET_STAT(tcp.rst),
+	       GET_STAT(tcp.rexmit));
+	printk("TCP conn drop  %d\tconnrst\t%d\n",
+	       GET_STAT(tcp.conndrop),
+	       GET_STAT(tcp.connrst));
+#endif
+
 #if defined(CONFIG_NET_STATISTICS_RPL)
 	printk("RPL DIS recv   %d\tsent\t%d\tdrop\t%d\n",
 	       GET_STAT(rpl.dis.recv),
diff --git a/subsys/net/ip/net_stats.c b/subsys/net/ip/net_stats.c
index 7b4f700a192..0a3aeacd45f 100644
--- a/subsys/net/ip/net_stats.c
+++ b/subsys/net/ip/net_stats.c
@@ -23,7 +23,7 @@ struct net_stats net_stats;
 
 #define PRINT_STATISTICS_INTERVAL (30 * MSEC_PER_SEC)
 
-void net_print_statistics(void)
+static inline void stats(void)
 {
 	static s64_t next_print;
 	s64_t curr = k_uptime_get();
@@ -80,6 +80,27 @@ void net_print_statistics(void)
 			 GET_STAT(udp.chkerr));
 #endif
 
+#if defined(CONFIG_NET_STATISTICS_TCP)
+		NET_INFO("TCP bytes recv %u\tsent\t%d",
+			 GET_STAT(tcp.bytes.received),
+			 GET_STAT(tcp.bytes.sent));
+		NET_INFO("TCP seg recv   %d\tsent\t%d\tdrop\t%d",
+			 GET_STAT(tcp.recv),
+			 GET_STAT(tcp.sent),
+			 GET_STAT(tcp.drop));
+		NET_INFO("TCP seg resent %d\tchkerr\t%d\tackerr\t%d",
+			 GET_STAT(tcp.resent),
+			 GET_STAT(tcp.chkerr),
+			 GET_STAT(tcp.ackerr));
+		NET_INFO("TCP seg rsterr %d\trst\t%d\tre-xmit\t%d",
+			 GET_STAT(tcp.rsterr),
+			 GET_STAT(tcp.rst),
+			 GET_STAT(tcp.rexmit));
+		NET_INFO("TCP conn drop  %d\tconnrst\t%d",
+			 GET_STAT(tcp.conndrop),
+			 GET_STAT(tcp.connrst));
+#endif
+
 #if defined(CONFIG_NET_STATISTICS_RPL)
 		NET_INFO("RPL DIS recv   %d\tsent\t%d\tdrop\t%d",
 			 GET_STAT(rpl.dis.recv),
@@ -129,6 +150,14 @@ void net_print_statistics(void)
 	}
 }
 
+void net_print_statistics(void)
+{
+	/* In order to make the info print lines shorter, use shorter
+	 * function name.
+	 */
+	stats();
+}
+
 #endif /* CONFIG_NET_STATISTICS_PERIODIC_OUTPUT */
 
 #if defined(CONFIG_NET_STATISTICS_USER_API)
diff --git a/subsys/net/ip/net_stats.h b/subsys/net/ip/net_stats.h
index a9d6cd07631..93b9669d620 100644
--- a/subsys/net/ip/net_stats.h
+++ b/subsys/net/ip/net_stats.h
@@ -9,6 +9,7 @@
 
 #if defined(CONFIG_NET_STATISTICS)
 
+#include <net/net_ip.h>
 #include <net/net_stats.h>
 
 extern struct net_stats net_stats;
@@ -162,6 +163,106 @@ static inline void net_stats_update_udp_drop(void)
 #define net_stats_update_udp_drop()
 #endif /* CONFIG_NET_STATISTICS_UDP */
 
+#if defined(CONFIG_NET_STATISTICS_TCP)
+/* TCP stats */
+static inline void net_stats_update_tcp_sent(u32_t bytes)
+{
+	net_stats.tcp.bytes.sent += bytes;
+}
+
+static inline void net_stats_update_tcp_recv(u32_t bytes)
+{
+	net_stats.tcp.bytes.received += bytes;
+}
+
+static inline void net_stats_update_tcp_resent(u32_t bytes)
+{
+	net_stats.tcp.resent += bytes;
+}
+
+static inline void net_stats_update_tcp_seg_sent(void)
+{
+	net_stats.tcp.sent++;
+}
+
+static inline void net_stats_update_tcp_seg_recv(void)
+{
+	net_stats.tcp.recv++;
+}
+
+static inline void net_stats_update_tcp_seg_drop(void)
+{
+	net_stats.tcp.drop++;
+}
+
+static inline void net_stats_update_tcp_seg_rst(void)
+{
+	net_stats.tcp.rst++;
+}
+
+static inline void net_stats_update_tcp_seg_conndrop(void)
+{
+	net_stats.tcp.conndrop++;
+}
+
+static inline void net_stats_update_tcp_seg_connrst(void)
+{
+	net_stats.tcp.connrst++;
+}
+
+static inline void net_stats_update_tcp_seg_chkerr(void)
+{
+	net_stats.tcp.chkerr++;
+}
+
+static inline void net_stats_update_tcp_seg_ackerr(void)
+{
+	net_stats.tcp.ackerr++;
+}
+
+static inline void net_stats_update_tcp_seg_rsterr(void)
+{
+	net_stats.tcp.rsterr++;
+}
+
+static inline void net_stats_update_tcp_seg_rexmit(void)
+{
+	net_stats.tcp.rexmit++;
+}
+#else
+#define net_stats_update_tcp_sent(...)
+#define net_stats_update_tcp_resent(...)
+#define net_stats_update_tcp_recv(...)
+#define net_stats_update_tcp_seg_sent()
+#define net_stats_update_tcp_seg_recv()
+#define net_stats_update_tcp_seg_drop()
+#define net_stats_update_tcp_seg_rst()
+#define net_stats_update_tcp_seg_conndrop()
+#define net_stats_update_tcp_seg_connrst()
+#define net_stats_update_tcp_seg_chkerr()
+#define net_stats_update_tcp_seg_ackerr()
+#define net_stats_update_tcp_seg_rsterr()
+#define net_stats_update_tcp_seg_rexmit()
+#endif /* CONFIG_NET_STATISTICS_TCP */
+
+static inline void net_stats_update_per_proto_recv(enum net_ip_protocol proto)
+{
+	if (IS_ENABLED(CONFIG_NET_UDP) && proto == IPPROTO_UDP) {
+		net_stats_update_udp_recv();
+	} else if (IS_ENABLED(CONFIG_NET_TCP) && proto == IPPROTO_TCP) {
+		net_stats_update_tcp_seg_recv();
+	}
+}
+
+static inline void net_stats_update_per_proto_drop(enum net_ip_protocol proto)
+{
+	if (IS_ENABLED(CONFIG_NET_UDP) && proto == IPPROTO_UDP) {
+		net_stats_update_udp_drop();
+	} else if (IS_ENABLED(CONFIG_NET_TCP) && proto == IPPROTO_TCP) {
+		net_stats_update_tcp_seg_drop();
+	}
+}
+
 #if defined(CONFIG_NET_STATISTICS_RPL)
 /* RPL stats */
 static inline void net_stats_update_rpl_resets(void)
diff --git a/subsys/net/ip/tcp.c b/subsys/net/ip/tcp.c
index 60cef5ae471..bff89bf5427 100644
--- a/subsys/net/ip/tcp.c
+++ b/subsys/net/ip/tcp.c
@@ -34,6 +34,7 @@
 #include "ipv6.h"
 #include "ipv4.h"
 #include "tcp.h"
+#include "net_stats.h"
 
 /*
  * Each TCP connection needs to be tracked by net_context, so
@@ -158,6 +159,11 @@ static void tcp_retry_expired(struct k_timer *timer)
 		do_ref_if_needed(pkt);
 		if (net_tcp_send_pkt(pkt) < 0 && !is_6lo_technology(pkt)) {
 			net_pkt_unref(pkt);
+		} else {
+			if (IS_ENABLED(CONFIG_NET_STATISTICS_TCP) &&
+			    !is_6lo_technology(pkt)) {
+				net_stats_update_tcp_seg_rexmit();
+			}
 		}
 	} else if (IS_ENABLED(CONFIG_NET_TCP_TIME_WAIT)) {
 		if (tcp->fin_sent && tcp->fin_rcvd) {
@@ -635,6 +641,8 @@ int net_tcp_queue_data(struct net_context *context, struct net_pkt *pkt)
 
 	context->tcp->send_seq += data_len;
 
+	net_stats_update_tcp_sent(data_len);
+
 	sys_slist_append(&context->tcp->sent_list, &pkt->sent_list);
 
 	/* We need to restart retry_timer if it is stopped. */
@@ -715,6 +723,8 @@ int net_tcp_send_pkt(struct net_pkt *pkt)
 			ret = net_send_data(new_pkt);
 			if (ret < 0) {
 				net_pkt_unref(new_pkt);
+			} else {
+				net_stats_update_tcp_seg_rexmit();
 			}
 
 			return ret;
@@ -773,6 +783,11 @@ void net_tcp_ack_received(struct net_context *ctx, u32_t ack)
 	u32_t seq;
 	bool valid_ack = false;
 
+	if (IS_ENABLED(CONFIG_NET_STATISTICS_TCP) &&
+	    sys_slist_is_empty(list)) {
+		net_stats_update_tcp_seg_ackerr();
+	}
+
 	while (!sys_slist_is_empty(list)) {
 		head = sys_slist_peek_head(list);
 		pkt = CONTAINER_OF(head, struct net_pkt, sent_list);
@@ -781,6 +796,7 @@ void net_tcp_ack_received(struct net_context *ctx, u32_t ack)
 		seq = sys_get_be32(tcphdr->seq) + net_pkt_appdatalen(pkt) - 1;
 
 		if (!net_tcp_seq_greater(ack, seq)) {
+			net_stats_update_tcp_seg_ackerr();
 			break;
 		}