The face recognition technique is considered as one of the successful applications in image perception. The current face recognition techniques have attained an adequate level, but their efficiency is confined by certain facts like illumination, pose changes and face expression. The proposed method uses Active Appearance Model (AAM) for face recognition. This technique is associated with computer vision and is used for shape and texture matching to a new image. Using optimization technique the estimated parameters and new images are matched. Existing 2D AAM methods fails because of change in face pose. To overcome this, 3D AAM is used in which the shape and appearance patterns are used to model human faces. During the training phase only front pose is required and using this data, nonfrontal face pose is successfully aligned. After alignment, Principal Component Analysis (PCA) is enforced along with the aligned model for texture variation.