Many studies have been conducted to analyse the influence diffusion in social networks, where the problem is normally modelled from a centralised perspective. In this paper, we propose an agent-based approach to model the bidirectional influence propagation in directed weighted networks in a distributed manner. In the proposed model, each user's personalised traits and individual's social context are taken into consideration on the basis of social theories. Moreover, the model can capture the dynamics in the environment as well. We claim that the proposed approach is more suitable for simulating the real-world complex influence propagation and predicting the evolution of entire network comparing with traditional approaches.