Abstract: A non-linear online model of China civil aviation risk was developed by using gray neural network method. Risk model inputs are represented by risk monitoring indicators about China civil aviation safety and output is represented by composite safety index of assessing China civil aviation industry safety. Numerical computation software was programmed based on this model. The agreement of the perception data of computation software with the actual data of composite safety index indicates that,using gray neural network method is correct. Using this risk model,the main influential factors from numerous risk monitoring indicators about civil aviation safety can be found out,and the reasonable proposal can be made for reducing China civil aviation safety risk. On the other hand,analysis of composite safety index of assessing China civil aviation industry safety can be made for providing a warning about China civil aviation safety.