Abstract: The object of this study concerns the determination of an optimal control of a nonlinear complex process. The proposed idea is based on the decomposition of the system into interconnected subsystems which are linear with respect to their state and control vectors. The nonlinearities are located in the interconnection terms. We have used the mixed method with two levels iterative computation structure. A fuzzy logic approach has been used in order to identify the optimal gains. The considered applications treat the optimal control of a rotary crane.