<p>To address functional rehabilitation needs following ankle injuries, an intelligent trajectory planning method for the combined movements of an ankle rehabilitation robot is proposed. Kinematic equations of the robot are derived by use of the vector method. Within the reachable workspace, a quintic closed B-spline trajectory is synthesized to combine sagittal-frontal ankle motions. In order to balance the rapid generation of trajectories with patient safety, a composite optimization function is presented, integrating four clinical indices: rehabilitation duration, patient comfort, actuator stationarity, and safety constraint. A trajectory planning method based on the chaotic opposition particle-enhanced sparrow search algorithm (CO-PSSA) is developed, which is composed of three strategies: chaotic opposition population initialization, hybrid producer strategy, and optimum-guided temporal repair strategy. Comprehensive simulations demonstrate that the proposed algorithm offers advantages in terms of rapid trajectory generation and enhanced safety. Additionally, trajectory tracking experiments are conducted on the physical prototype, and the results validate that the planned trajectory exhibits smooth transitions and high reliability.</p>