Trust is very essential to secure and high quality interactions on the Semantic Web. In this paper, we introduce a trust model, which incorporates pairwise trust ratings and reliable factors of acquaintances and constructs an edge-weighted graph to calculate trust values. We interpret trust in two dimensions to better deal with it. We apply the theory in social network field to calculate trust and simulate the trust relationships between humans. We also take advantage of formulas in probability and statistics to analyze the similarities between agents. The algorithm of the trust model aims to be simple, efficient and flexible.