In this paper, a disturbance-observer–based control approach is developed for overhead crane systems. Different from existing control strategies, the issues consisting of the output feedback, input saturation, double-pendulum dynamics, and uncertain disturbances are taken into consideration here. In particular, a disturbance observer is designed first, which can exactly estimate uncertain disturbances. Next, to enhance the performance of the controller, a virtual position signal is constructed and a corresponding Lyapunov function is introduced. Then, based on the provided Lyapunov function and the designed disturbance observer, a composite control approach is developed for overhead crane systems with double-pendulum dynamics and the convergence of the system states is proved via rigorous theoretical analysis. Finally, the effectiveness and robustness of the proposed control approach are verified by simulation tests.