Person identification and recognition using fingerprints as a biometric authentication is one of the unique, prominent and reliable method of verification processes. The paper presents fingerprint authentication system using filter bank based matching algorithm. The algorithm uses Gabor filter bank to extract features of fingerprints. These features are then compared for identification and recognition of a person. The Euclidian distance is used to compare feature vector of query image with feature vectors of database fingerprints. Two sets of databases FVC2000 and DBIT are used for evaluating the performance and efficiency of the algorithm. The experimental results show that the average efficiency of the algorithm is 82.95% and 89.68% for FVC2000 and DBIT fingerprint databases respectively.