In order to tackle the robot trajectory planning problem with the short running time as the optimization goal, a time-optimal trajectory planning algorithm was presented based on improved simplified particle swarm optimization (ISPSO). The robot's trajectory was constructed by 3-5-3 polynomial interpolation in the joint space of the robot. Under the condition of satisfying the velocity constraint, the objective function was constructed by the sum of the time intervals between each node. ISPSO was used to optimize the objective function. The algorithm was improved by optimizing the inertia weight updating method and introducing a golden sine segmentation algorithm as an optimization operator. Compared with other particle swarm optimization algorithms, ISPSO had higher search velocity and accuracy. The effectiveness of the proposed algorithm was demonstrated through simulations using the PUMA 560 industrial robot, which resulted in a 19% reduction in time compared to the simplified particle swarm algorithm. The simulation results show that ISPSO achieved time optimization under the condition of velocity constraint, which proved its superiority in trajectory planning.