Phosphorylation plays a very important role in the post translational modification, because it determines the enzyme and receptor if get together or not. That indirectly affects signal transmission and immune system in our body. Although the technology of identify phosphorylation develop little by little, the amount of phosphorylation data still not many. Therefore, it is necessary for the prediction of phosphorylation. In this study, we fuse the predictive data which was got from the three predictive websites (Netphos, Disphose and GPS) to predict phosphorylation sites. The voting only mechanism also considered in our final test. Finally, we got a good evaluative value than other methods by testing results.