Abstract: Traditional researches of pairwise influence quantification are focus on social network such as coauthor network.So they don't fully apply to current online social network.Furthermore,existing model which are proposed for online social network are simple,one-sided and lack of accuracy.So we propose a Bayes model which integrated all useful information online social network provides to quantify the pairwise influence between two users.Through analyzing and comparing the influential found by these models,we find that our model is more accurate and comprehensive than existing models.