diff --git a/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/dropout_gpu_kernel.h b/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/dropout_gpu_kernel.h
index 632caef9ec..1fe3cff151 100644
--- a/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/dropout_gpu_kernel.h
+++ b/mindspore/ccsrc/backend/kernel_compiler/gpu/nn/dropout_gpu_kernel.h
@@ -54,12 +54,18 @@ class DropoutGpuFwdKernel : public GpuKernel {
     float *mask_f = GetDeviceAddress<float>(workspace, 0);
 
     if (!states_init_) {
-      curandCreateGenerator(&mask_generator_, CURAND_RNG_PSEUDO_DEFAULT);
-      curandSetPseudoRandomGeneratorSeed(mask_generator_, time(NULL));
+      CHECK_CURAND_RET_WITH_EXCEPT(curandCreateGenerator(&mask_generator_, CURAND_RNG_PSEUDO_DEFAULT),
+                                   "Failed to create generator");
+      CHECK_CURAND_RET_WITH_EXCEPT(curandSetPseudoRandomGeneratorSeed(mask_generator_, time(NULL)),
+                                   "Failed to SetPseudoRandomGeneratorSeed");
+      MS_EXCEPTION_IF_NULL(mask_generator_);
       states_init_ = true;
     }
+    CHECK_CURAND_RET_WITH_EXCEPT(curandSetStream(mask_generator_, reinterpret_cast<cudaStream_t>(stream_ptr)),
+                                 "Failed to set stream for generator");
     // curandGen only support float or double for mask.
-    curandGenerateUniform(mask_generator_, mask_f, num_count_);
+    CHECK_CURAND_RET_WITH_EXCEPT(curandGenerateUniform(mask_generator_, mask_f, num_count_),
+                                 "Failed to generate uniform");
     DropoutForward(input, mask, output, mask_f, num_count_, keep_prob_, reinterpret_cast<cudaStream_t>(stream_ptr));
 
     return true;
diff --git a/mindspore/ccsrc/runtime/device/gpu/gpu_common.h b/mindspore/ccsrc/runtime/device/gpu/gpu_common.h
index ea2b321714..83f00e0fae 100644
--- a/mindspore/ccsrc/runtime/device/gpu/gpu_common.h
+++ b/mindspore/ccsrc/runtime/device/gpu/gpu_common.h
@@ -20,7 +20,9 @@
 #include <iostream>
 #include <vector>
 #include <algorithm>
+#include <map>
 #include "utils/log_adapter.h"
+#include "include/curand.h"
 
 namespace mindspore {
 namespace device {
@@ -131,6 +133,15 @@ inline bool CheckNullInput(std::vector<size_t> input_shape) {
   return false;
 }
 #define CHECK_NULL_INPUT(input_shape) mindspore::device::gpu::CheckNullInput(input_shape)
+
+#define CHECK_CURAND_RET_WITH_EXCEPT(expression, message)                                     \
+  {                                                                                           \
+    curandStatus_t status = (expression);                                                     \
+    if (status != CURAND_STATUS_SUCCESS) {                                                    \
+      MS_LOG(EXCEPTION) << "CUAD curand Error: " << message << " | curandStatus: " << status; \
+    }                                                                                         \
+  }
+
 }  // namespace gpu
 }  // namespace device
 }  // namespace mindspore
diff --git a/tests/st/ops/gpu/test_dropout.py b/tests/st/ops/gpu/test_dropout.py
new file mode 100644
index 0000000000..9a9f9d5b09
--- /dev/null
+++ b/tests/st/ops/gpu/test_dropout.py
@@ -0,0 +1,54 @@
+# Copyright 2020 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+import numpy as np
+import pytest
+
+import mindspore.nn as nn
+from mindspore import Tensor
+from mindspore.ops import operations as P
+
+
+class Net(nn.Cell):
+    def __init__(self, keep_prob):
+        super(Net, self).__init__()
+        self.drop = P.Dropout(keep_prob)
+
+    def construct(self, x_):
+        return self.drop(x_)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_gpu_training
+@pytest.mark.env_onecard
+def test_dropout():
+    x_shape = [32, 16, 2, 5]
+    x = np.ones(x_shape).astype(np.float32)
+    keep_prob = 0.4
+    dropout = Net(keep_prob)
+    tx = Tensor(x)
+    output, mask = dropout(tx)
+    # check output
+    output_np = output.asnumpy()
+    elem_count = x.size
+    nonzero_count = np.count_nonzero(output_np)
+    assert (elem_count * (keep_prob - 0.1)) < nonzero_count < (elem_count * (keep_prob + 0.1))
+    output_sum = np.sum(output_np)
+    x_sum = np.sum(x)
+    assert abs(output_sum - x_sum)/x_sum < 0.1
+    # check mask
+    mask_np = mask.asnumpy()
+    mask_sum = np.sum(mask_np)
+    assert np.count_nonzero(mask_np) == nonzero_count
+    assert abs(mask_sum - nonzero_count)/nonzero_count < 0.1