The three-dimensional feeding manipulator is studied to improve working efficiency after meeting the requirements of the operation,and the trajectory planning and optimization with the optimal execution time are put forward. The trajectory is planned by B-spline curve and the execution time of manipulator working is optimized by multi-parent genetic algorithm. The optimization model of shortest time of the manipulator is established once a time,which takes joint angular velocity,joint acceleration and joint impact as the optimal constraint. To test the model,some simulation and experiment are conducted by 11 key trajectory points,and the simulation result show that the optimization method can improve the working efficiency under the premise of satisfying the job requirement. The algorithm is verified with test-bed,it shows that the experiment result is consistence with the simulation result. This method has a significance of the study on the trajectory planning of the three dimensional feeding manipulator.