diff --git a/PaddleCV/image_classification/README.md b/PaddleCV/image_classification/README.md
index 896720eafccc0d5bec49a4c2df41417ebcddeef8..83ff508a6b0bd9aa12b43541b49dfca6a7d06d33 100644
--- a/PaddleCV/image_classification/README.md
+++ b/PaddleCV/image_classification/README.md
@@ -46,7 +46,8 @@ pip install numpy
 
 ### 数据准备
 
-下面给出了ImageNet分类任务的样例，首先，通过如下的方式进行数据的准备：
+下面给出了ImageNet分类任务的样例
+在Linux系统下通过如下的方式进行数据的准备：
 ```
 cd data/ILSVRC2012/
 sh download_imagenet2012.sh
@@ -69,6 +70,8 @@ val/ILSVRC2012_val_00000001.jpeg 65
 ```
 注意：可能需要根据本地环境调整reader.py中相关路径来正确读取数据。
 
+**Windows系统下请用户自行下载ImageNet数据。[label下载链接](http://paddle-imagenet-models.bj.bcebos.com/ImageNet_label.tgz)**
+
 ### 模型训练
 
 数据准备完毕后，可以通过如下的方式启动训练：
diff --git a/PaddleCV/image_classification/README_en.md b/PaddleCV/image_classification/README_en.md
index f9b68ca3b18f2a1470e3b7bc7d9243ea064ea11b..18a073ae2c071af931321551e0211d8d969b74fc 100644
--- a/PaddleCV/image_classification/README_en.md
+++ b/PaddleCV/image_classification/README_en.md
@@ -37,7 +37,8 @@ Running samples in this directory requires Python 2.7 and later, CUDA 8.0 and la
 
 ### Data preparation
 
-An example for ImageNet classification is as follows. First of all, preparation of imagenet data can be done as:
+An example for ImageNet classification is as follows.
+For Linux system, preparation of imagenet data can be done as:
 
 ```bash
 cd data/ILSVRC2012/
@@ -62,6 +63,7 @@ val/ILSVRC2012_val_00000001.jpeg 65
 ```
 
 Note: You may need to modify the data path in reader.py to load data correctly.
+**For windows system, Users should download ImageNet data by themselves. and the label list can be downloaded in [Here](http://paddle-imagenet-models.bj.bcebos.com/ImageNet_label.tgz)**
 
 ### Training