Aiming at the problem that structural similarity (SSIM) method has different sensitivity to different types of image distortion, an image quality evaluation method based on local variance and structural similarity is proposed. Calculate and compare the global local variance distribution of the two images, and analyze the image characteristics represented by the measurement results. Local variance better reflects the detailed information of the image, so it is more sensitive to distortion types such as blur, but less sensitive to distortion types such as noise pollution. Therefore, the evaluation method proposed in this paper combines local variance and structural similarity methods to balance their different sensitive deviations to various factors that cause image distortion, so that the evaluation results are closer to the subjective perception of human eyes on image quality. The experimental results show that this method fully considers the influence of the integrity of human sensitive details on image quality, and the evaluation results are better than SSIM method, which accords with human subjective feelings.