An automated fingerprint verification system has become a focus in the field of pattern recognition in the world with the development of image processing,pattern recognition,and computer technique since 1960s,particularly the fingerprint matching that is one of the most important problems in the fingerprint verification system.A fingerprint verification based on fuzzy discrimination is proposed in this paper which can modify the minutiae matching algorithm put forward by Jain et al,and better solve the problem of detecting only a number of false minutiae and losing many true ones when the image quality is very poor.This algorithm can be capable of noise resistance and nonlinear deformation.