Abstract: The problem of pattern recognition based on characteristic vector is often encountered in computer measurement and control system. But the characteristic vectors of existing patterns in database and that of object will be recognized are all to present somewhat uncertainty due to the error from the procedure of extracting them. This research puts fuzzy theory into recognizing procedure with transforming it into a similarity or distance problem of two fuzzy sets. The recognition algorithms based on shortest distance and largest grid similarity degree are provided in this paper, and to be successfully used in personal identification system based on hand shape.