Taking the inland water navigation safety as the research object, analyze the factors influencing the shipping safety state with the viewpoint of systems engineering from the four aspects: crew, vessel, environment and management. Then, establish the Bayesian network model with the factors as the node to evaluate shipping safety state. Finally, Construct the Bayesian network model with the date of the marine accident and risk on Yangtze River for evidence reasoning by software GeNIe, which can measure the action mechanism of the factors in causing risk and also reflect the navigation safety status. The safety index system includes comprehensive index and subindex. Comprehensive index reflect the navigation safety trend with its change scope, and the sub-index represent the main risk factors in shipping safety. Safety index system can quantify the inland water shipping safety state continuously and dynamically, which could makes the safety monitoring more accurate.