Abstract: Aiming at the problem of low precision and slow speed of traditional algorithms, this paper proposes an improved SURF algorithm. The algorithm uses a brand new Harr-like feature descriptor to increase the number of valid feature points and improve the discriminant ability of the descriptors. At the same time we combined with BBF search method to enhance the matching speed. The experimental results show that the matching point of the improved algorithm is increased by more than 15% compared with the original SURF algorithm, and the speed is also improved. The improvement of the algorithm opens up a broader prospect for the application of image matching technology in the printing industry.