In this paper, an efficient Face recognition system based on Haar wavelet and Block Independent Component Analysis (BICA) algorithm is presented. The Adaptive boosting classifier is trained with Haar features to classify and to detect face. To recognize the detected face, the dimensionality reduced approach of Independent Component analysis; BICA is used. This system is capable of detecting the presence of faces in the field of view of camera in real time and verify the person's identity if the image of the person is already been trained by the system. The system is able to detect the face even with slight occlusions and under varying illumination conditions, improving the recognition significantly.