diff --git a/deploy/python/README.md b/deploy/python/README.md
index 4b28f0cf83da7d6195aad574ccdd5a3560d614eb..cb32ad468b58ff1d0cf0f9cb5f625822c8959393 100644
--- a/deploy/python/README.md
+++ b/deploy/python/README.md
@@ -91,6 +91,8 @@ python deploy/python/mot_keypoint_unite_infer.py --mot_model_dir=output_inferenc
 | --enable_mkldnn | Option | CPU预测中是否开启MKLDNN加速，默认为False |
 | --cpu_threads | Option| 设置cpu线程数，默认为1 |
 | --trt_calib_mode | Option| TensorRT是否使用校准功能，默认为False。使用TensorRT的int8功能时，需设置为True，使用PaddleSlim量化后的模型时需要设置为False |
+| --save_results | Option| 是否在文件夹下将图片的预测结果以JSON的形式保存 |
+
 
 说明：
 
diff --git a/deploy/python/infer.py b/deploy/python/infer.py
index 09eec1153796ae43f41f0881db22029ed6056e45..fd43874c3d883a39d68a435ee43667b18b254d1c 100644
--- a/deploy/python/infer.py
+++ b/deploy/python/infer.py
@@ -15,6 +15,8 @@
 import os
 import yaml
 import glob
+import json
+from pathlib import Path
 from functools import reduce
 
 import cv2
@@ -233,7 +235,8 @@ class Detector(object):
                       image_list,
                       run_benchmark=False,
                       repeats=1,
-                      visual=True):
+                      visual=True,
+                      save_file=None):
         batch_loop_cnt = math.ceil(float(len(image_list)) / self.batch_size)
         results = []
         for i in range(batch_loop_cnt):
@@ -293,6 +296,10 @@ class Detector(object):
             if visual:
                 print('Test iter {}'.format(i))
 
+        if save_file is not None:
+            Path(self.output_dir).mkdir(exist_ok=True)
+            self.format_coco_results(image_list, results, save_file=save_file)
+
         results = self.merge_batch_result(results)
         return results
 
@@ -313,7 +320,7 @@ class Detector(object):
         if not os.path.exists(self.output_dir):
             os.makedirs(self.output_dir)
         out_path = os.path.join(self.output_dir, video_out_name)
-        fourcc = cv2.VideoWriter_fourcc(* 'mp4v')
+        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
         writer = cv2.VideoWriter(out_path, fourcc, fps, (width, height))
         index = 1
         while (1):
@@ -337,6 +344,68 @@ class Detector(object):
                     break
         writer.release()
 
+    @staticmethod
+    def format_coco_results(image_list, results, save_file=None):
+        coco_results = []
+        image_id = 0
+
+        for result in results:
+            start_idx = 0
+            for box_num in result['boxes_num']:
+                idx_slice = slice(start_idx, start_idx + box_num)
+                start_idx += box_num
+
+                image_file = image_list[image_id]
+                image_id += 1
+
+                if 'boxes' in result:
+                    boxes = result['boxes'][idx_slice, :]
+                    per_result = [
+                        {
+                            'image_file': image_file,
+                            'bbox':
+                            [box[2], box[3], box[4] - box[2],
+                             box[5] - box[3]],  # xyxy -> xywh
+                            'score': box[1],
+                            'category_id': int(box[0]),
+                        } for k, box in enumerate(boxes.tolist())
+                    ]
+
+                elif 'segm' in result:
+                    import pycocotools.mask as mask_util
+
+                    scores = result['score'][idx_slice].tolist()
+                    category_ids = result['label'][idx_slice].tolist()
+                    segms = result['segm'][idx_slice, :]
+                    rles = [
+                        mask_util.encode(
+                            np.array(
+                                mask[:, :, np.newaxis],
+                                dtype=np.uint8,
+                                order='F'))[0] for mask in segms
+                    ]
+                    for rle in rles:
+                        rle['counts'] = rle['counts'].decode('utf-8')
+
+                    per_result = [{
+                        'image_file': image_file,
+                        'segmentation': rle,
+                        'score': scores[k],
+                        'category_id': category_ids[k],
+                    } for k, rle in enumerate(rles)]
+
+                else:
+                    raise RuntimeError('')
+
+                # per_result = [item for item in per_result if item['score'] > threshold]
+                coco_results.extend(per_result)
+
+        if save_file:
+            with open(os.path.join(save_file), 'w') as f:
+                json.dump(coco_results, f)
+
+        return coco_results
+
 
 class DetectorSOLOv2(Detector):
     """
@@ -807,7 +876,10 @@ def main():
         if FLAGS.image_dir is None and FLAGS.image_file is not None:
             assert FLAGS.batch_size == 1, "batch_size should be 1, when image_file is not None"
         img_list = get_test_images(FLAGS.image_dir, FLAGS.image_file)
-        detector.predict_image(img_list, FLAGS.run_benchmark, repeats=100)
+        save_file = os.path.join(FLAGS.output_dir,
+                                 'results.json') if FLAGS.save_results else None
+        detector.predict_image(
+            img_list, FLAGS.run_benchmark, repeats=100, save_file=save_file)
         if not FLAGS.run_benchmark:
             detector.det_times.info(average=True)
         else:
diff --git a/deploy/python/utils.py b/deploy/python/utils.py
index c542f0176494e03312516574077815fbdd2d6d4c..41dc7ae9e81f49bdd08e0917d50b21ac00f2e527 100644
--- a/deploy/python/utils.py
+++ b/deploy/python/utils.py
@@ -156,6 +156,12 @@ def argsparser():
         type=ast.literal_eval,
         default=False,
         help="Whether do random padding for action recognition.")
+    parser.add_argument(
+        "--save_results",
+        type=bool,
+        default=False,
+        help="Whether save detection result to file using coco format")
+
     return parser