非线性系统预测控制已成为一个有着广阔前景的研究热点。预测控制的多步预测和滚动优化有克服建模误差影响的能力，使控制系统具有一定的鲁棒性。但研究表明,预测控制对模型失配的鲁棒性是有限的，仍需要一个较准确的预测模型-Nonlinear predictive control has become a research focus has broad prospects. Predictive control multi-step prediction and rolling optimization have overcome modeling errors affect the ability of the control system has a certain robustness. But studies show that the robustness of the model predictive control of mismatch is limited, still we need a more accurate prediction model