In this paper, we give a framework to deblur the blurry frame in a video clip. Two kinds of motion blurring effects can be removed in the video, one is the blurring effect caused by hand shaking, the other is the blurring effect caused by a fast moving object. For the blurring caused by hand shaking, PSF is estimated by comparing the stable area in blurry frame and non-blurry frame, so the Richardson-Lucy algorithm can restore the blurry frame by non-blind deconvolution. We also propose a framework to deblur the motion blurring objects which move fast in the video. The background is reconstructed by the algorithm in each frame, so an accurate matte of blurry object can be extracted to deblur the moving object by alpha matting. Results show our method is effective.