diff --git a/python/paddle/tests/test_vision_models.py b/python/paddle/tests/test_vision_models.py
index 5f35a1e0e5a4ba9d6c6683918a2a03c190089762..a25a8f373c29c4e678f87453eade1fd958c9ac33 100644
--- a/python/paddle/tests/test_vision_models.py
+++ b/python/paddle/tests/test_vision_models.py
@@ -71,6 +71,9 @@ class TestVisonModels(unittest.TestCase):
     def test_resnet152(self):
         self.models_infer('resnet152')
 
+    def test_vgg16_num_classes(self):
+        vgg16 = models.__dict__['vgg16'](pretrained=False, num_classes=10)
+
     def test_lenet(self):
         input = InputSpec([None, 1, 28, 28], 'float32', 'x')
         lenet = paddle.Model(models.__dict__['LeNet'](), input)
diff --git a/python/paddle/vision/models/vgg.py b/python/paddle/vision/models/vgg.py
index bb158569d3bc9fb658c1e103b41bb02784e68d8b..00f6cccbdfe9f11c0f234e923383ecc566958a33 100644
--- a/python/paddle/vision/models/vgg.py
+++ b/python/paddle/vision/models/vgg.py
@@ -107,10 +107,7 @@ cfgs = {
 
 
 def _vgg(arch, cfg, batch_norm, pretrained, **kwargs):
-    model = VGG(make_layers(
-        cfgs[cfg], batch_norm=batch_norm),
-                num_classes=1000,
-                **kwargs)
+    model = VGG(make_layers(cfgs[cfg], batch_norm=batch_norm), **kwargs)
 
     if pretrained:
         assert arch in model_urls, "{} model do not have a pretrained model now, you should set pretrained=False".format(