Abstract: A new nonlinear model predictive control (NMPC) scheme to guarantee closed-loop exponential stability is presented for Hammerstein systems subject to constraints on the state and input. Solving errors of the nonlinear algebra equation are predicted recedingly based on stabilized and optimal control laws of the linear subsystem. And then predictive control actions are computed to satisfy the constraint via online optimization. Moreover, the bound of solving errors is obtained to ensure the exponential stability of the closed-loop system. Hence, the closed-loop system not only satisfies the constraint but also has robustness property with respect to the solving error. Finally, a simulation example of industrial polypropylene grade transition control is used to illustrate the effectiveness of the algorithm proposed here.