diff --git a/paddle/fluid/operators/shape_op_npu.cc b/paddle/fluid/operators/shape_op_npu.cc
new file mode 100644
index 0000000000000..fa6cecf818aff
--- /dev/null
+++ b/paddle/fluid/operators/shape_op_npu.cc
@@ -0,0 +1,61 @@
+/* Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+
+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. */
+
+#include <memory>
+#include <string>
+
+#include "paddle/fluid/operators/npu_op_runner.h"
+#include "paddle/fluid/operators/shape_op.h"
+
+namespace paddle {
+namespace operators {
+
+using Tensor = framework::Tensor;
+using SelectedRows = framework::SelectedRows;
+
+template <typename DeviceContext, typename T>
+class ShapeNPUKernel : public framework::OpKernel<T> {
+ public:
+  void Compute(const framework::ExecutionContext& ctx) const override {
+    auto* in_var = ctx.InputVar("Input");
+    framework::DDim in_dims;
+    if (in_var->IsType<SelectedRows>()) {
+      in_dims = in_var->Get<SelectedRows>().value().dims();
+    } else {
+      in_dims = in_var->Get<LoDTensor>().dims();
+    }
+    auto* out_t = ctx.Output<Tensor>("Out");
+    out_t->Resize({in_dims.size()});
+    // to do: cpuplace?
+    auto out_data = out_t->mutable_data<int32_t>(platform::CPUPlace());
+    for (int i = 0; i < in_dims.size(); ++i) {
+      out_data[i] = in_dims[i];
+    }
+  }
+};
+
+}  // namespace operators
+}  // namespace paddle
+
+namespace ops = paddle::operators;
+
+REGISTER_OP_NPU_KERNEL(
+    shape, ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, bool>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, int>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, int8_t>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, uint8_t>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, int64_t>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, float>,
+    ops::ShapeNPUKernel<paddle::platform::NPUDeviceContext, double>);
+#endif
diff --git a/python/paddle/fluid/tests/unittests/npu/test_shape_op_npu.py b/python/paddle/fluid/tests/unittests/npu/test_shape_op_npu.py
new file mode 100644
index 0000000000000..7b9a74b2be98d
--- /dev/null
+++ b/python/paddle/fluid/tests/unittests/npu/test_shape_op_npu.py
@@ -0,0 +1,57 @@
+#  Copyright (c) 2021 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+
+from __future__ import print_function
+
+import numpy as np
+import unittest
+import sys
+sys.path.append("..")
+from op_test import OpTest
+import paddle
+import paddle.fluid as fluid
+
+paddle.enable_static()
+SEED = 2021
+
+
+@unittest.skipIf(not paddle.is_compiled_with_npu(),
+                 "core is not compiled with NPU")
+class TestShape(OpTest):
+    def setUp(self):
+        self.set_npu()
+        self.op_type = "shape"
+        self.place = paddle.NPUPlace(0)
+
+        self.init_dtype()
+        np.random.seed(SEED)
+        x = np.random.uniform(1, 2, [5, 10]).astype(self.dtype)
+        out = np.array([5, 10])
+
+        self.inputs = {'Input': OpTest.np_dtype_to_fluid_dtype(x)}
+        self.attrs = {}
+        self.outputs = {'Out': out}
+
+    def set_npu(self):
+        self.__class__.use_npu = True
+
+    def init_dtype(self):
+        self.dtype = np.float32
+
+    def test_check_output(self):
+        self.check_output_with_place(self.place, check_dygraph=False)
+
+
+if __name__ == '__main__':
+    unittest.main()