Abstract: This paper proposes a fifth-degree polynomial-based position control strategy to address the crane robot impact problem in the start-stop phase of large-load lifting equipment, as well as the defects of traditional automatic positioning devices that rely on visual pre-testing and have low braking accuracy. During the speed change process, This strategy enables smooth speed changes in segments under dynamic load by generating a time-displacement curve and establishing a fully parameterized acceleration-velocity-displacement coupling model. During the stopping process, upon receiving the in-place signal, this strategy can immediately reconstruct the remaining path trajectory, guide the device to stop accurately without repeated fine-tuning, and significantly shorten the positioning time. The continuity of the algorithm's speed change process and the accuracy of parking have been verified through simulations and experiments. This method has been a reliable solution for high-precision motion control in unmanned scenarios. It has significant application value.