The prediction accuracy and generalization ability of the support vector regression(SVR) model depend on a proper setting of its parameters to a great extent. An optimal selection approach of SVR parameters is proposed based on the multi-agent particle swarm optimization(MAPSO) algorithm. On this basis, a model predictive control method based on the MAPSO-SVR is proposed and applied to the nonlinear predictive control scheme to select the optimal control inputs. For the nonlinear system, the simulation results show that the proposed method is effective and has an excellent adaptive ability and robustness. Compared with the model predictive controllers based on SVR optimized by particle swarm optimization algorithm(PSO-SVR), SVR optimized genetic algorithm(GA-SVR), and RBF neural network algorithm, the proposed method is superior to other methods.