To exclude the interference of clouds to scanning laser fuze, meet the requirements of identification precision and real-time performance of fuze when missile and target intersect in high speed, a target feature recognition method based on the characteristics of corner point was proposed. The method uses MATLAB to extract corner point of target matrix and calculate the rectangle degrees of corner point area and use corner point and rectangle degrees of corner point area to distinguish between target and clouds, and improve the real-time performance of aircraft target identification. Three-dimensional model was used to simulate aircraft flight attitude which verified Harris algorithm with variety of attitude of the local plane image matrix. The simulation results show that using this method, target and clouds can be quickly and efficiently distinguished.