In order to optimize the work of robot end-hand trajectory, improve working efficiency, reduce energy loss, established the optimization model with the goal of shortest exercise time and path, minimum impact. Based on traditional ant colony algorithm established the overall inspiration factor with the introduction of information to improve the optimal of path searching efficiency and used the death of ants mechanisms and penalty function to avoid the pitfalls encountered in the path formed at the deadlock situation. Test results show that the improved ant colony algorithm is faster convergence and can map out the optimal path in a relatively short period of time to meet the conditions.