The feature point-based matching methods suffer from unsatisfactory matching effect when employed in complex texture scenes. Therefore, a block based image matching method, which combines Speeded Up Robust Features(SURF) algorithm and Coherency-Sensitive Hashing(CSH) is proposed. This method is divided into three stages: firstly, using the SURF algorithm to extract feature points from the image; secondly, determining the feature areas with the SURF feature points as the center; thirdly, the feature areas are matched by CSH. In order to further speed up the method, SURF algorithm is modified to remove the calculation of the direction of feature points when extracting feature points. Experimental results show that the new method is better than the general image matching methods based on feature point in complex texture image matching.The efficiency of the proposed method has been increased by 10%to 15%compared with that of the CSH algorithm.