In practical robot control systems, the input constraints problem, caused by a limited actuator output amplitude, is due to its own physical characteristics. In traditional control methods, this nonlinear problem is not paid enough attention, especially for robot systems with unknown disturbance. To handle this problem, adaptive compensate control for uncertain robot system with input constraints is proposed. It is combined with adaptive compensate control and hyperbolic tangent function to ensure that the uncertain robot system with input constraints is stable. The simulation results showed that the proposed method can make the uncertain robot system reach a stable state.