Aiming at the problem of long running time and low matching accuracy in image matching process，random sample consensus ( ＲANSAC) algorithm is used to optimize the speed-up of robust features ( SUＲF)optimization algorithm. As a result，an adaptable algorithm is proposed to optimize image matching. Firstly，theSUＲF operator is used for feature detection and feature description. Then the neighbor algorithm is used to prematchthe feature points. Finally，the random sampling consistency ( ＲANSAC) algorithm is used to optimize thematching results. The scale invariant feature transform ( SIFT) algorithm，SUＲF algorithm，and the proposedoptimization algorithm are compared in the same experimental environment. Compared with the SIFT algorithm andthe SUＲF algorithm，the optimization algorithm reduces the number of matching point pairs to 38 and 18 pairs，excluding mismatched points，improving the matching accuracy，and reducing the running time of the algorithm.