Abstract: In order to analyze the stability of slopes, by using coordinated particle swarm optimization(CPSO) and BP neural network, a CPSO-BP model is built to predict slope stability. BP neural network can describe the complex nonlinear relationship between the slope stability and its influential factors; six main influential factors, i.e. internal friction angle, slope angle, unit weight of rock, slope height, cohesion and pore pressure ratio, as a network input, the network output is the slope stability factor. BP neural network is so easy to fall into local optimum that take advantage of coordinated particle swarm optimization algorithm the ability of global to determine BP neural network connection weights and thresholds, let the advantages of BP neural network can fully play. Frinally we will approach to the aim that improving the accuracy of model predictions. Case study shows that the CPSO-BP model has better prediction accuracy and it is feasible to predict the slope stability.

Key words: particle swarm optimization(PSO), coordination, BP neural network, slope, stability evaluation