Abstract: As the complexity of external factors and internal mechanics in slope，there is a highly non-linear relativity between slope stability and its influence factors，it is hard to construct a general model to describe the process of change of slope stability，SVM(Support Vectors Machine)is a intelligent experience learning algorithm，it performs great in processing the nonlinear problem.Analyzed the slope factor data with the SVM algorithm then got the experience learning model about corresponding relativity between slope stability and its influence factors，finally did the evaluation with the experience model get previously.Experience results showed that the method proposed performed well in evaluating state of the slope stability；it just could satisfy the need of projects.

Key words: Support vector machine, Parameter optimization, Pattern recognition, Slope, Stability evaluation