Abstract: In this paper,a improved BP neural network model was used for evaluating the stability of a rock slope has been built by use of the NNtool of MATLAB.Tested the reliability of the neural network model by an analysis of a project in Tianjin .Chose some parameters about rock integrity as the input factors base on the actual situation,collected some rock slope engineering cases to train the model and prove the neural network model is suitable. From tested the results which export from computer,this model is proved to be feasible and it is wider application.