Person re-identification is to match the same person’s images captured at different space and different time across non-overlapping camera views. Person re-identification has important application value in video surveillance and is an important basis for intelligent surveillance systems. However, due to factors such as low camera performance, monitoring scene changes and pedestrian mobility, at this stage person re-identification has many challenges such as low image resolution, illumination,background, viewpoint and pose changes. This paper analyzes that the direct removal of the background brings new gradient information at the boundary, which makes the performance of the person re-identification model worse. In this regard, it proposes a simple Gaussian weighting based background suppression method, which can effectively suppress the background and improve the performance of person recognition model.