Abstract: The micro level characteristics of the innovation diffusion process in a small world social network were studied by using agentbased complex system simulation method. The effect of Nashrule and Paretorule on the diffusion process and diffusion result was analyzed. The simulation indicates that the payoff value of the agent is higher when the agent uses Pareto rule when the cost factor k is within a certain range． Furthermore, the average payoff value of the agents has an inflexion with the changing of k, and the inflexion appears at k=3.7.