In this paper, we apply augmented reality technology to make human computer interface more technically advanced and interesting. We present a real-time face tracking system for augmented reality, which will be used in electronic commerce, to help shoppers acquire a more direct interaction with the products they purchase. A real-time face tracking mechanism can enhance the realism of online shopping. In order to emphasize the convenience and practicability of our system, we used plug-in functions and Principal Component Analysis (PCA) to conduct real-time face tracking and Neural Networks (NN) to reduce training time and achieve valid recognition. Convenience and uniqueness are the other main parts of this system.