The work aims to solve the problems of maximum speed constraint, short running time, and poor operation smoothness of palletizing robots when grasping target objects. With a Func200id six-axis industrial robot as the research object, its DH parameters were calculated by the relationship of each joint size of the robot, and a simulation model was established. The improved HS algorithm changed the search strategy of the algorithm by increasing the variation factor VF so that each individual in the algorithm not only had the opportunity to participate in the iterative process but also had the chance to learn from the optimal individual, which accelerated the convergence of the algorithm. Through a comparison experiment of the four algorithms running 30 times under six different test functions, the improved HS algorithm had stronger search capability and faster convergence than the remaining three algorithms. It can significantly shorten the working hours of the palletizing robot and run continuously and smoothly, which has certain application value in other industrial scenarios as well.