Abstract: Imaging and detection of the targets behind the wall are usually interfered by strong wall reflections in Through­the­Wall Radar Imaging(TWRI)．The targets near the wall may even be drowned completely.Therefore，mitigating the wall clutter is of great significance for identifying the target clearly in radar imaging.Traditional methods for wall­clutter mitigation in single input and single output system(SISO)usually make use of interference canceling and spatial filtering，which are based on the invariance of the spatial characteristics of wall.The spatial filter is applied along the array aperture to remove zero­frequency and low­frequency components，which correspond to the wall electromagnetic responses.However，they are not applicable to multiple inputs and multiple outputs systems(MIMO)．To remove the clutter for UWB MIMO TWR，this paper presents an effective SVDmethod based on subspace projection.As an Eigen­structure technique，SVD is applied to the received signal matrix to extract the target signatures.The signal space is decomposed into three subspaces：the clutter subspace，the target subspace，and the noise subspace.The clutter subspace mainly includes wall reflections，which is much stronger than target signal and noise.The clutter subspace and the target subspace can be isolated from the signal matrix according to the dominant singular components(SCs)．And then，orthogonal subspace projection are used to mitigate wall components and make target signals prominent.Finally，MATLAB simulation software is used for modeling and simulating.From the experimental results，we can see that the proposed method is effective in removing wall reflections，suppressing clutter，and highlighting the target.This method is superior to interference canceling method.Furthermore，this SVDmethod can also remove random noises and suppress false imaging in multi­object detecting，improving image quality and reliability.