The performance of a character recognition system depends heavily on what features are being used. In this paper, an optical character recognition method based on match of feature lines is presented. By extracting the feature lines from bitmap of character and detecting the necessary-sufficient condition with templates by baseline superposed, this method carries out the recognition with a very high efficiency. The process of selecting and adjusting the feature lines from template is also described. Then, a new way for collating is proposed. By the hand of dispalaying skeletons of recognized characters that overlap the original bitmap,this method makes finding errors easier.Finally,the performance evaluation based on comparison is given.