Fingerprint matching is the most accurate method among the biometrics. They are used for recognizing the person identity. The matching of two fingerprint images based on the numbering of minutiae and bifurcation points. Minutiae based Algorithm only extract local features of fingerprint image. The new algorithm used pattern based method which covers the whole area of fingerprint image. It includes both local and global features of fingerprint image. This new algorithm divides in three phases. First phase include enhancing the image quality with the thresholding value and reducing the dimensional space of fingerprint image with principal component analysis. Principal Component analysis contains the Principal components in lower dimensionality space with high value information. In second phase, Gabor filter is used for extracting feature vectors from fingerprint image and number of feature vectors form feature map. In third phase, Euclidean distance is calculated between two feature maps to determine whether the two fingerprint images are same or not. The result evaluation is done on the basis of bifurcation and minutiae points and their comparisons with previous work.