A novel facial expression recognition technique is proposed in this thesis. It can further be used in distance learning. First, it employs the face detection method to locate the face region in images, which integrate the Haar-like feature and the self quotient image filter to improve the detection rate due to the insufficient light and shade light. Second, face-block normalization is also carried out. Subsequently, angular radiual transform, discrete cosine transform and Gabor filter are used in the procedure of facial expression feature extraction. A trained support vector machine is conducted to predict the expression for an query face block. Finally, experimental results show that the performance of the proposed technique can be better than that of other methods.