In online social network, the personal information dissemination behavior is reported to be affected by the clash of social individuals' biased opinions. In this paper, we present a model to discuss the influence of individual biased opinion on diffusion dynamics. Based on multi-agent simulations, we obtain some conclusions which are helpful for recommender systems and in controlling diffusion. In addition, our study offers potential avenues for the study of diffusion dynamics with personal biases.