Abstract: The problem of license plate recognition in inclement weather. Traditional PCA vehicle video detection methods, when in the bad outside weather like in poor light conditions, there is loss of image pixels, leading to the problem can not accurately identify the license plate. In order to improve the accuracy of license plate recognition, this paper presents a movement license plate recognition algorithm. By the three image sequences adjacent to the vertical and horizontal directions for the poor projection, a full segmentation of images of vehicles to avoid the read only a single pixel to solve the problem of missing pixels to achieve an accurate license plate recognitioru Experimental results show that using this algorithm the adverse weather conditions on the vehicle license plate image segmentation and complete the correct identification, to obtain satisfactory results.