Image restoration technique is an important method to improve the quality of image in flights photograph and machine vision. On account of blur emerge in relative motion and other condition, the paper analyse its mathematic principle, then for each motion pattern we adopt different deblurring method. Be concrete, we first judge the type of motion, for linear motion blur we take both hough transform and error-parameter to estimate the blur parameters, yet for rotate motion blur we take curve-fitting and polar transformation to estimate parameters. After that we apply Wiener filtering to recover image and improve image quality. We perform experiments with linear motion blur parameter (30, 70) and rotate motion blur parameter ((128,128), 20°). The contrast results of the experiments indicate that the algorithm can enhance image quality effectively. Compared with traditional handling, ours have the advantage of being high precision, robustness and speediness..