<p>In this paper, the adaptive optimal control problem is investigated for human-in-the-loop (HiTL) dual-arm robot systems with actuator faults under the privacy preservation mechanism. Firstly, in order to prevent the leakage of sensitive information, an improved user-defined time mask function is constructed to achieve privacy preservation in the setting time, and the initial value of the dual-arm robot system is manually adjusted without the influence of the system order and the protected time. When the trajectory of the reference signal changes suddenly due to the external environment, the human operators will issue a decision signal according to the actual physical significance of the parameters, and then the output signal converges to the trajectory guided by the dynamic reference signal. Based on a simplified reinforcement learning (RL) framework, the controller of the dual-arm robot system is optimized to improve the performance of the system. Meanwhile, the proposed control scheme can guarantee that all the signals are semiglobally uniformly ultimately bounded, and the tracking errors can converge to a small neighborhood of the origin. Finally, a simulation example is presented to demonstrate the effectiveness of the proposed scheme.</p>