With the expansion of the world, the risk management of finance and financial supply chain has paid more and more attention to the risk of a single enterprise or industry. To this end, this project intends to combine Markov decision making and Bayesian network to study new ideas of risk management in financial and financial markets. On this basis, a probability distribution model based on probability distribution is studied and combined with past and future prediction results to achieve better prediction results. Bayesian model is used to correlate different risk factors, so as to provide a comprehensive basis for the policy-making of relevant departments. Experimental results show that compared with single Markov decision or Bayesian network, this method can improve the precision of risk identification, the effect of assistant decision and the effectiveness of risk response decision, and the recognition accuracy is up to 93.66%. Through the above research, the integration model improves the overall risk control level of financial supply chain, so as to improve the overall risk control level of financial supply chain. This project can study the risk of financial supply chain deeply, which has great academic significance and application value.