In order to prevent Internet financial risks and guarantee the role of financial auditing in the risk management of Internet financial platform operations, based on the bounded rationality hypothesis, this paper uses evolutionary game theory to analyze the decision-making behavior of audit institutions and Internet financial platforms in supervision and operation. It introduces the “scenario-response” thought of computational experiments to simulate the evolutionary game results. The experimental results show that under the current Internet finance ecology in China, the illegal operation of Internet financial platform is a stable and balanced strategy. By adjusting the behavioral returns, losses and cost parameters of the audit institutions, the strategic choice probability of strict audit supervision can be improved, and then it is proposed to improve the reward and punishment mechanism of audit supervision, optimize the cost of audit supervision and establish a long-term Internet financial audit and supervision system.