In this paper, the problem of the motion control of flexible manipulators with disturbances and system uncertainties is studied in multiple time scales. Due to the multi time scale characteristics of flexible manipulator, it is easy to lead to ill conditioned numerical problems when using traditional methods to design controllers. Based on singular perturbation theory, the flexible manipulator is modeled as a singular perturbation system with singular perturbation parameters. And then it is decomposed into slow and fast subsystems which describe rigid motion and flexible vibration separately. In the slow time scale, considering the influence of external disturbance and parameter uncertainty, an adaptive sliding mode controller (slow controller) with a disturbance observer is designed to track the trajectory of the flexible manipulator. In the fast time scale, in order to eliminate the influence of system uncertainties such as un-modeled dynamics and vibration in slow time scale on the controller design, a robust sliding mode controller (fast controller) is designed to suppress the vibration of the flexible manipulator. Finally, the singular perturbation theory is used to combine the slow and fast controller to achieve the dual control objectives of tracking the desired position while suppressing the vibration. The experimental results show that the method proposed in this paper is robust and effective.