Aiming at the problem that the image corner features matching accuracy of the traditional ORB algorithm is not high， an image corner features matching method based on optimized ORB algorithm is proposed. Firstly， this paper uses the Shi-Tomasi algorithm to detect the image corner features， then uses the fusion algorithm of BRIEF and SURF to generate the double-descriptor sequences of the image corner features， and reduces dimension of the double-descriptor sequences by the random projection principle， finally， the optimized matching algorithm is used to match， referred to as the Shi-Tomasi-SURFORB algorithm. The simulation experiment analyzes seven indicators， including the number of corner features， the number of corner features matching， the number of correct corner features matching， the accuracy of corner features matching， the corner feature matching accuracy ratio， the image matching time， and the image matching time rate， the analysis results show that compared with the traditional ORB algorithm， the Shi-Tomasi-SURFORB algorithm improves the image corner features matching time by 9.46%， but improves the image corner features matching accuracy by 8.88%， which provides a more balanced solution for image corner features matching.