Abstract This paper presents a hybrid algorithm for solving complex optimal control problems based on decomposition. The general finite-time optimal control problem for a class of hybrid dynamical systems is considered, which has not been solved in a decomposed way by existing methods. The problem is first decomposed into a master problem and a subproblem, and then the two are linked via logic-based Benders decomposition. Computational experiments have been carried out for the considered problem. The results show that the proposed algorithm could substantially reduce the solving time, compared with directly solving by mixed integer solvers.