diff --git a/test_tipc/configs/yolov3/yolov3_darknet53_270e_coco_model_linux_gpu_normal_normal_serving_python_linux_gpu_cpu.txt b/test_tipc/configs/yolov3/yolov3_darknet53_270e_coco_model_linux_gpu_normal_normal_serving_python_linux_gpu_cpu.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0a07de5246525b5014eb46564eeab09a034f7293
--- /dev/null
+++ b/test_tipc/configs/yolov3/yolov3_darknet53_270e_coco_model_linux_gpu_normal_normal_serving_python_linux_gpu_cpu.txt
@@ -0,0 +1,20 @@
+===========================cpp_infer_params===========================
+model_name:yolov3_darknet53_270e_coco
+python:python
+filename:null
+##
+--output_dir:./output_inference
+weights:https://paddledet.bj.bcebos.com/models/yolov3_darknet53_270e_coco.pdparams
+norm_export:tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml -o
+quant_export:tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml --slim_config configs/slim/quant/yolov3_darknet_qat.yml -o
+fpgm_export:tools/export_model.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml --slim_config configs/slim/prune/yolov3_darknet_prune_fpgm.yml -o
+distill_export:null
+export1:null
+export2:null
+kl_quant_export:tools/post_quant.py -c configs/yolov3/yolov3_darknet53_270e_coco.yml --slim_config configs/slim/post_quant/yolov3_darknet53_ptq.yml -o
+--export_serving_model:True
+##
+start_serving:-m paddle_serving_server.serve --model serving_server 
+--port:9393 
+--gpu_ids:0
+##
diff --git a/test_tipc/prepare.sh b/test_tipc/prepare.sh
index c706e7069b95ace483f40a361b9fd35dab76b260..754d5c24c92247d3ecf87c3aeacda565bca2b64c 100644
--- a/test_tipc/prepare.sh
+++ b/test_tipc/prepare.sh
@@ -71,6 +71,16 @@ elif [ ${MODE} = "paddle2onnx_infer" ];then
     # set paddle2onnx_infer enve
     ${python} -m pip install install paddle2onnx
     ${python} -m pip install onnxruntime==1.10.0
+elif [ ${MODE} = "serving_infer" ];then
+    git clone https://github.com/PaddlePaddle/Serving
+    bash Serving/tools/paddle_env_install.sh
+    cd Serving
+    pip install -r python/requirements.txt
+    cd ..
+    pip install paddle-serving-client==0.8.3 -i https://pypi.tuna.tsinghua.edu.cn/simple
+    pip install paddle-serving-app==0.8.3 -i https://pypi.tuna.tsinghua.edu.cn/simple
+    pip install paddle-serving-server-gpu==0.8.3.post101 -i https://pypi.tuna.tsinghua.edu.cn/simple
+    python -m pip install paddlepaddle-gpu==2.2.2.post101 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html
 else
     # download coco lite data
     wget -nc -P ./dataset/coco/ https://paddledet.bj.bcebos.com/data/tipc/coco_tipc.tar
diff --git a/test_tipc/test_serving.sh b/test_tipc/test_serving.sh
index e69de29bb2d1d6434b8b29ae775ad8c2e48c5391..8c122d96dffa4437e16199f6772a027d8b984d42 100644
--- a/test_tipc/test_serving.sh
+++ b/test_tipc/test_serving.sh
@@ -0,0 +1,63 @@
+#!/bin/bash
+source test_tipc/utils_func.sh
+
+FILENAME=$1
+
+# parser model_name
+dataline=$(cat ${FILENAME})
+IFS=$'\n'
+lines=(${dataline})
+model_name=$(func_parser_value "${lines[1]}")
+echo "ppdet serving: ${model_name}"
+python=$(func_parser_value "${lines[2]}")
+filename_key=$(func_parser_key "${lines[3]}")
+filename_value=$(func_parser_value "${lines[3]}")
+
+# export params
+save_export_key=$(func_parser_key "${lines[5]}")
+save_export_value=$(func_parser_value "${lines[5]}")
+export_weight_key=$(func_parser_key "${lines[6]}")
+export_weight_value=$(func_parser_value "${lines[6]}")
+norm_export=$(func_parser_value "${lines[7]}")
+pact_export=$(func_parser_value "${lines[8]}")
+fpgm_export=$(func_parser_value "${lines[9]}")
+distill_export=$(func_parser_value "${lines[10]}")
+export_key1=$(func_parser_key "${lines[11]}")
+export_value1=$(func_parser_value "${lines[11]}")
+export_key2=$(func_parser_key "${lines[12]}")
+export_value2=$(func_parser_value "${lines[12]}")
+kl_quant_export=$(func_parser_value "${lines[13]}")
+export_serving_model_key=$(func_parser_key "${lines[14]}")
+export_serving_model_value=$(func_parser_value "${lines[14]}")
+# parser serving
+start_serving=$(func_parser_value "${lines[16]}")
+port_key=$(func_parser_key "${lines[17]}")
+port_value=$(func_parser_value "${lines[17]}")
+gpu_id_key=$(func_parser_key "${lines[18]}")
+gpu_id_value=$(func_parser_value "${lines[18]}")
+
+LOG_PATH="./test_tipc/output"
+mkdir -p ${LOG_PATH}
+status_log="${LOG_PATH}/results_serving.log"
+
+function func_serving(){
+    IFS='|'
+    if [ ${gpu_id_key} = "null" ]; then
+        start_serving_command="nohup ${python} ${start_serving} ${port_key} ${port_value} > serving.log 2>&1 &"
+    else
+        start_serving_command="nohup ${python} ${start_serving} ${port_key} ${port_value} ${gpu_id_key} ${gpu_id_value} > serving.log 2>&1 &"
+    fi
+    echo $start_serving_command
+    eval $start_serving_command
+    last_status=${PIPESTATUS[0]}
+    status_check $last_status "${start_serving_command}" "${status_log}"
+}
+cd output_inference/${model_name}
+echo $PWD
+func_serving
+test_command="${python} ../../deploy/serving/test_client.py ../../deploy/serving/label_list.txt ../../demo/000000014439.jpg"
+echo $test_command
+eval $test_command
+last_status=${PIPESTATUS[0]}
+status_check $last_status"${test_command}" "${status_log}"
+