Fingerprint matching is one of the most important problems in AFIS. In general, we use minutiae such as ridge endings and ridge bifurcation to represent a fingerprint and do fingerprint matching through minutiae matching. In this paper, we proposed a minutia-matching algorithm, which modified Jain et al.'s algorithm. Our algorithm can better distinguish two images from different fingers and is more robust to nonlinear deformation. Experiments done on a set of fingerprint images captured with an inkless scanner shows that our algorithm is fast and has high accuracy.