Face recognition (FR) system can be applied in security measure at air ports, passport verification, criminals list, verify police department, visa processing, verification of electoral identification and card Security measure at automated teller machine (ATM). Difficulties inherent in human face recognition from unconstrained and motion characterized scenes are usually accounted for, due to varying illumination and pose of subjects in the scenes. However, a number of technical approaches have been developed to manage these nuisance factors to make recognition possible and optimal. In this paper, we proposed a FR system using PCA and KPCA for face recognition. The image used in the database is one hundred and fifty five (155) altogether, ninety (90) for training and sixty five (65) for testing. The images were collected in an uncontrolled environment with different pose and expressions. A total number of one hundred and fifty five images (155) was used to test the effectiveness and compatibility of the developed system with a real time facial expression which were taken at varying tilt angles. The proposed system obtained good recognition accuracy but it varies for the two different classifications, the system was able to obtain 92% accuracy for PCA which is lesser than that of KPCA which got a result of 96% accuracy. From the results obtained, it was observed that based on the classification, KPCA outperforms PCA in terms of the accuracy.