Abstract:This research built a flight phase safety risk assessment model basing on Back Propagation(BP) neural network. By using MATLAB software and flight data of A320，the training and verification of the model were realized，and the errors were acceptable. The results shows that the model is capable of predicting the possible risk in a flight phase. Moreover，the flight phase safety risk assessment system can provide scientific solutions and can be taken as a tool, which facilitates not only the management regarding flight safety for an airline company，but also the role in flight supervision for the civil aviation administration.