The work aims to propose a character correction method based on polynomial self-optimizing algorithm regarding the problems of no prior knowledge of the distortion, low efficiency of traditional method and poor accuracy in the correction process of distortion print characters. The character area to be detected was treated as a small image. The distortion region was determined by the initial correction. With the minimum difference result between the image to be tested and the standard image as the optimization target, the self-optimizing operation of the correction control point was carried out in distortion region by bacterial population algorithm, thereby establishing a correction function to correct the character distortion region, and finally it was realized by Matlab simulation. The mean correction accuracy of this method was less than 0.6 pixels, and the running time was less than 0.12 s, which achieved the purpose of fast and accurate correction of distortion characters. The self-optimizing method has overcome the shortcomings of manual operation and effectively corrected the distorted characters, which helps to improve the accuracy and efficiency of subsequent defect detection.