Abstract: An identification method for nonlinear systems using support vector machine is investigated. And a kind of nonlinear predictive control method based on support vector machine model is presented. The simulation result for a CSTR process shows that the support vector machine is of good ability of modeling nonlinear process under small data set available and is of good generalization. The nonlinear predictive control strategy based on support vector machine model shows satisfactory performance. Thus this predictive control strategy introduces a new way for generic nonlinear control.