In this paper, we propose an enhanced QoS prediction approach to predict the missing QoS values for QoS-based selection. Compared with existing QoS prediction methods, our proposed A Smooth approach has two differences: 1) using A-cosine equation for similarity calculation to remove the impact of different QoS scale, 2) adding a data Smoothing process to improve the prediction accuracy. An extensive performance study based on a real public dataset is reported to verify the prediction accuracy of our proposed approach.