In order to reduce the flight operation risk caused by flight operation controllers’individual differences, and further improve the airlines operation control ability, based on the systemic analysis of flight operation procedures, combined with the operation data, screening of risk factors from the aspects of flight crew, airports, weather, routes and aircraft, the flight operation control risk assessment index is established. Using the Bayesian network analysis method based on the fault tree and insecurity event analysis reports over the years as the sample data, the probability of occurrence of unsafe events by forward reasoning is predicted. Adopting three types of importance index analysis to identify the key risk factors. The result shows that five key risk factors probability value are more than 40% resulting in a high flight operation risk value, such as pilot and co- pilot’s technical level, crew matching and air traffic controller cooperation, etc. Risk inference result is entirely consistent with actual operation situation．
