By introducing information fusion techniques into optimal preview control, a new theory of Information Fusion Optimal Preview Control (IFOPC) for nonlinear discrete system is proposed in this paper. All information on control system which includes ideal control strategy, expected objective trajectory and dynamics of system, is regarded as measure information of control system. Therefore, the problem of optimal control can be transferred into the problem of information fusion estimate. It should be pointed out that the IFOPC consists of optimal co-state generator and optimal regulator, in which the optimal co-state generator is just a extended Kalman filter. Firstly, the information fusion estimate theorems are described in brief. Secondly, an algorithm of IFOPC is deduced in detail. Thirdly, some key techniques of IFOPC are discussed, which include the selection of preview step number and the determination of Kalman filter initial values. Finally, the validity of the proposed control strategy is confirmed by the simulation of manipulator shift control. The simulation results show good tracking performance of the IFOPC based nonlinear system.