diff --git a/src/backends/backendsCommon/test/CMakeLists.txt b/src/backends/backendsCommon/test/CMakeLists.txt
index ea214de..bfaca6c 100644
--- a/src/backends/backendsCommon/test/CMakeLists.txt
+++ b/src/backends/backendsCommon/test/CMakeLists.txt
@@ -73,6 +73,8 @@
     layerTests/ConvertBf16ToFp32TestImpl.hpp
     layerTests/ConvertFp16ToFp32TestImpl.cpp
     layerTests/ConvertFp16ToFp32TestImpl.hpp
+    layerTests/ConvertFp32ToBf16TestImpl.cpp
+    layerTests/ConvertFp32ToBf16TestImpl.hpp
     layerTests/ConvertFp32ToFp16TestImpl.cpp
     layerTests/ConvertFp32ToFp16TestImpl.hpp
     layerTests/DebugTestImpl.cpp
diff --git a/src/backends/backendsCommon/test/IsLayerSupportedTestImpl.hpp b/src/backends/backendsCommon/test/IsLayerSupportedTestImpl.hpp
index a070ac0..d646847 100644
--- a/src/backends/backendsCommon/test/IsLayerSupportedTestImpl.hpp
+++ b/src/backends/backendsCommon/test/IsLayerSupportedTestImpl.hpp
@@ -447,6 +447,8 @@
 
 DECLARE_LAYER_POLICY_1_PARAM(ConvertFp16ToFp32)
 
+DECLARE_LAYER_POLICY_1_PARAM(ConvertFp32ToBf16)
+
 DECLARE_LAYER_POLICY_1_PARAM(ConvertFp32ToFp16)
 
 DECLARE_LAYER_POLICY_2_PARAM(Convolution2d)
diff --git a/src/backends/backendsCommon/test/LayerTests.hpp b/src/backends/backendsCommon/test/LayerTests.hpp
index 1c6277a..600a261 100644
--- a/src/backends/backendsCommon/test/LayerTests.hpp
+++ b/src/backends/backendsCommon/test/LayerTests.hpp
@@ -15,6 +15,7 @@
 #include <backendsCommon/test/layerTests/ConcatTestImpl.hpp>
 #include <backendsCommon/test/layerTests/ConvertBf16ToFp32TestImpl.hpp>
 #include <backendsCommon/test/layerTests/ConvertFp16ToFp32TestImpl.hpp>
+#include <backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.hpp>
 #include <backendsCommon/test/layerTests/ConvertFp32ToFp16TestImpl.hpp>
 #include <backendsCommon/test/layerTests/Conv2dTestImpl.hpp>
 #include <backendsCommon/test/layerTests/ConstantTestImpl.hpp>
diff --git a/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.cpp b/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.cpp
new file mode 100644
index 0000000..66eb4ee
--- /dev/null
+++ b/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.cpp
@@ -0,0 +1,77 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#include "ConvertFp32ToBf16TestImpl.hpp"
+
+#include <backendsCommon/test/TensorCopyUtils.hpp>
+#include <backendsCommon/test/WorkloadTestUtils.hpp>
+
+#include <test/TensorHelpers.hpp>
+
+LayerTestResult<armnn::BFloat16, 4> ConvertFp32ToBf16Test(
+    armnn::IWorkloadFactory& workloadFactory,
+    const armnn::IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
+{
+    IgnoreUnused(memoryManager);
+
+    const armnn::TensorInfo inputTensorInfo({1, 2, 4, 3}, armnn::DataType::Float32);
+    const armnn::TensorInfo outputTensorInfo({1, 2, 4, 3}, armnn::DataType::BFloat16);
+
+    auto input = MakeTensor<float, 4>(inputTensorInfo,
+        { -37.5f, -15.2f, -8.76f,
+          -2.0f, -1.5f, -1.3f,
+          -0.5f, -0.4f, 0.0f,
+          1.0f, 0.4f, 0.5f,
+          1.3f, 1.5f, 2.0f,
+          8.76f, 15.2f, 37.5f,
+          3.8f, // 0x40733333 Round down
+          3.1055E+29f, // 0x707ADC3C Round up
+          9.149516E-10f, // 0x307B7FFF Round down
+          -3.8f, // 0xC0733333 Round down
+          -3.1055E+29f, // 0xF07ADC3C Round up
+          -9.149516E-10f // 0xB07B7FFF Round down
+        });
+
+    std::vector<armnn::BFloat16> outputValues = armnnUtils::QuantizedVector<armnn::BFloat16>(
+        {
+            -37.5f, -15.2f, -8.76f,
+            -2.0f, -1.5f, -1.3f,
+            -0.5f, -0.4f, 0.0f,
+          1.0f, 0.4f, 0.5f,
+          1.3f, 1.5f, 2.0f,
+          8.76f, 15.2f, 37.5f,
+          3.796875f, // 0x4073
+          3.1072295E29f, // 0x707B
+          9.131327E-10f, // 0x307B
+          -3.796875f, // 0xC073
+          -3.1072295E29f, // 0xF07B
+          -9.131327E-10f // 0xB07B
+        },
+        1.0f, 0);
+
+    LayerTestResult<armnn::BFloat16, 4> ret(outputTensorInfo);
+    ret.outputExpected = MakeTensor<armnn::BFloat16, 4>(outputTensorInfo, outputValues);
+
+    std::unique_ptr<armnn::ITensorHandle> inputHandle = workloadFactory.CreateTensorHandle(inputTensorInfo);
+    std::unique_ptr<armnn::ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
+
+    armnn::ConvertFp32ToBf16QueueDescriptor data;
+    armnn::WorkloadInfo info;
+    AddInputToWorkload(data, info, inputTensorInfo, inputHandle.get());
+    AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
+
+    std::unique_ptr<armnn::IWorkload> workload = workloadFactory.CreateConvertFp32ToBf16(data, info);
+
+    inputHandle->Allocate();
+    outputHandle->Allocate();
+
+    CopyDataToITensorHandle(inputHandle.get(), &input[0][0][0][0]);
+
+    workload->Execute();
+
+    CopyDataFromITensorHandle(&ret.output[0][0][0][0], outputHandle.get());
+
+    return ret;
+}
diff --git a/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.hpp b/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.hpp
new file mode 100644
index 0000000..4c6125f
--- /dev/null
+++ b/src/backends/backendsCommon/test/layerTests/ConvertFp32ToBf16TestImpl.hpp
@@ -0,0 +1,17 @@
+//
+// Copyright © 2020 Arm Ltd. All rights reserved.
+// SPDX-License-Identifier: MIT
+//
+
+#pragma once
+
+#include "LayerTestResult.hpp"
+
+#include <BFloat16.hpp>
+
+#include <armnn/backends/IBackendInternal.hpp>
+#include <backendsCommon/WorkloadFactory.hpp>
+
+LayerTestResult<armnn::BFloat16, 4> ConvertFp32ToBf16Test(
+    armnn::IWorkloadFactory& workloadFactory,
+    const armnn::IBackendInternal::IMemoryManagerSharedPtr& memoryManager);