Abstract: The RBF(Radial Basis Function)neural networks model for rock slope stability evaluation is built,which eliminates the slowly convergence speed of BP(Back Propagation)networks and gives an effective quantified method for evaluating the rock slope stability.A new method for designing and optimizing the open-pit slope based on RBF neural networks is proposed,which is applied to designing and optimizing ShuiChang Iron Mine open-pit slope angle.