Abstract: In the lunar surface sampling mission, the lunar sampling manipulator, due to the flexibility of its boom and joints, has low stiffness, which leads to the reduction of manipulator control accuracy. To address the above problem, in this paper, an end-operation error compensation method was proposed based on the estimation of flexible deformation of the manipulator, and an overall stiffness model of the multi-degree-of-freedom manipulator was constructed by analyzing joint flexibility, arm rod flexibility and self-weight and other factors on joint flexibility of the end position of the manipulator. The error bounds for the end position estimation of the manipulator under different configuration conditions were given. On this basis, a one-time compensation control method for the end operation of the manipulator was designed, which can reduce the maximum control deviation of the end of the robotic arm from 35 mmto less than 1 mm, greatly improving the absolute positioning accuracy of the lunar sampling manipulator simulation system.