diff --git a/misc/tools/acrntrace/Makefile b/misc/tools/acrntrace/Makefile
index 22ff4f901..c44532e31 100644
--- a/misc/tools/acrntrace/Makefile
+++ b/misc/tools/acrntrace/Makefile
@@ -11,6 +11,7 @@ TRACE_CFLAGS += -Werror
 TRACE_CFLAGS += -O2 -U_FORTIFY_SOURCE -D_FORTIFY_SOURCE=2
 TRACE_CFLAGS += -Wformat -Wformat-security -fno-strict-aliasing
 TRACE_CFLAGS += -fpie -fpic
+TRACE_CFLAGS += -lnuma
 TRACE_CFLAGS += $(CFLAGS)
 
 GCC_MAJOR=$(shell echo __GNUC__ | $(CC) -E -x c - | tail -n 1)
diff --git a/misc/tools/acrntrace/README.rst b/misc/tools/acrntrace/README.rst
index b56e7c2b9..786427f8f 100644
--- a/misc/tools/acrntrace/README.rst
+++ b/misc/tools/acrntrace/README.rst
@@ -24,6 +24,7 @@ Options:
 -i period               specify polling interval in milliseconds [1-999]
 -t max_time             max time to capture trace data (in second)
 -c                      clear the buffered old data
+-a cpu-set              only capture the trace data on these configured cpu-set
 
 acrntrace_format.py
 ===================
diff --git a/misc/tools/acrntrace/acrntrace.c b/misc/tools/acrntrace/acrntrace.c
index 03078a44b..2cbdef985 100644
--- a/misc/tools/acrntrace/acrntrace.c
+++ b/misc/tools/acrntrace/acrntrace.c
@@ -19,6 +19,7 @@
 #include <pthread.h>
 #include <string.h>
 #include <signal.h>
+#include <numa.h>
 
 #include "acrntrace.h"
 
@@ -29,7 +30,7 @@ static int exiting = 0;
 
 /* for opt */
 static uint64_t period = 10000;
-static const char optString[] = "i:hct:";
+static const char optString[] = "i:hct:a:";
 static const char dev_prefix[] = "acrn_trace_";
 
 static uint32_t flags;
@@ -38,6 +39,8 @@ static char trace_file_dir[TRACE_FILE_DIR_LEN];
 static reader_struct *reader;
 static int dev_cnt = 0; /* Count of /dev/acrn_trace_xxx devices */
 
+static struct bitmask *cpu_bitmask = NULL;
+
 static void display_usage(void)
 {
 	printf("acrntrace - tool to collect ACRN trace data\n"
@@ -46,7 +49,8 @@ static void display_usage(void)
 	       "\t-h: print this message\n"
 	       "\t-i: period_in_ms: specify polling interval [1-999]\n"
 	       "\t-t: max time to capture trace data (in second)\n"
-	       "\t-c: clear the buffered old data\n");
+	       "\t-c: clear the buffered old data\n"
+	       "\t-a: cpu-set: only capture the trace data on these configured cpu-set\n");
 }
 
 static void timer_handler(union sigval sv)
@@ -116,6 +120,9 @@ static int parse_opt(int argc, char *argv[])
 		case 'c':
 			flags |= FLAG_CLEAR_BUF;
 			break;
+		case 'a':
+			cpu_bitmask = numa_parse_cpustring_all(optarg);
+			break;
 		case 'h':
 			display_usage();
 			return -EINVAL;
@@ -316,8 +323,10 @@ static void handle_on_exit(void)
 
 	pr_info("exiting - to release resources...\n");
 
-	foreach_dev(dev_id)
-	    destory_reader(&reader[dev_id]);
+	foreach_dev(dev_id) {
+		if (numa_bitmask_isbitset(cpu_bitmask, dev_id))
+			destory_reader(&reader[dev_id]);
+	}
 }
 
 static void signal_exit_handler(int sig)
@@ -341,6 +350,14 @@ int main(int argc, char *argv[])
 		exit(EXIT_FAILURE);
 	}
 
+	/* if we don't set the -a option or set it by mistake, capture the trace on all possible dev */
+	if (!cpu_bitmask) {
+		/* numa_bitmask_alloc will cause the process exiting on failure */
+		cpu_bitmask = numa_bitmask_alloc(dev_cnt);
+		foreach_dev(dev_id)
+			numa_bitmask_setbit(cpu_bitmask, dev_id);
+	}
+
 	reader = calloc(1, sizeof(reader_struct) * dev_cnt);
 	if (!reader) {
 		pr_err("Failed to allocate reader memory\n");
@@ -366,9 +383,11 @@ int main(int argc, char *argv[])
 
 	/* acquair res for each trace dev */
 	flags |= FLAG_TO_REL;
-	foreach_dev(dev_id)
-	    if (create_reader(&reader[dev_id], dev_id) < 0)
-		goto out_free;
+	foreach_dev(dev_id) {
+		if (numa_bitmask_isbitset(cpu_bitmask, dev_id))
+			if (create_reader(&reader[dev_id], dev_id) < 0)
+				goto out_free;
+	}
 
 	/* for kill exit handling */
 	signal(SIGTERM, signal_exit_handler);
@@ -380,8 +399,10 @@ int main(int argc, char *argv[])
 		printf("q <enter> to quit:\n");
 
  out_free:
-	foreach_dev(dev_id)
-	    destory_reader(&reader[dev_id]);
+	foreach_dev(dev_id) {
+		if (numa_bitmask_isbitset(cpu_bitmask, dev_id))
+			destory_reader(&reader[dev_id]);
+	}
 
 	free(reader);
 	flags &= ~FLAG_TO_REL;