This paper proposes an iterative learning control problem based on the differential evolution algorithm for optimal control gains. The proposed framework for a nonlinear discrete-time system consists of open-loop ILC component and closed-loop control component, forming an open-closed-loop ILC structure. The inclusion of the open-loop component guarantees the convergence of the ILC tracking error in terms of mathematical expectation. Feedback control accelerates convergence with appropriate gain. The control gain is optimized by the differential evolution algorithm to achieve better system control and faster convergence. After conducting ILC convergence analysis and simulation, the tracking error tends to approach zero in mathematical expectation.