The face images matching algorithm based on KAZE is to do nonlinear diffusion filtering by the additive operator splitting algorithm. In this way, the problem of blurred boundaries and detail missing can be solved. A stable nonlinear scale space is constructed by using arbitrary step to search the Hessian local maximum value point after different scales normalizing to detect feature points. By using M-SURE to describe the feature points, the feature vectors are constructed. The KAZE and SIFT feature are used to do face images matching under VS2010 and Opencv. By changing the blur level, angle of rotation, scale, change of brightness, a further simulation experiment can be conducted aiming at KAZE, SIFT, SURF in Matlab. The research result proves that the KAZE has better performance even if under the condition of Gaussian Blur, angle rotating, scale transformation and intensity roughness.