This paper proposes a laser cleaning detection method for rust surface of steel based on the machine vision technology. The laser cleaning visual detection system is constructed, and the SURF algorithm is used to continuously splice the collected images. Based on the HSV color space conversion method, the HSV threshold of steel rust is obtained, and the median filtering method is used to remove the interference pixels. The results show that the laser cleaning detection method based on machine vision can realize the visual detection of steel rust surface during the laser cleaning process. In comparison with the gray threshold extraction method, the detection accuracy is improved, which can provide quantitative evaluation for the laser cleaning quality of steel rust surface.