To address the issue of reduced production efficiency caused by robotic arms taking too long to perform complex, repetitive motions in industrial production, a multi-strategy improved sand cat swarm optimization (MISCSO) algorithm was proposed to achieve time-optimal trajectory planning for manipulators. The algorithm includes four improvement strategies: the population initialization strategy combining Tent chaotic mapping and refraction opposition-based learning, dynamic nonlinear sensitivity strategy, adaptive spiral search strategy, and sand cat ambush-raid predation strategy. Four strategies have significantly improved the solution accuracy and convergence performance of the algorithm. Taking a six-axis manipulator as the research object, the kinematic model was established by the D-H parameter method, and the joint trajectory was constructed based on the 3-5-3 polynomial interpolation function. The motion time of the manipulators was significantly optimized from 12 s to 8.663 1 s by using the MISCSO. Compared with the sand cat swarm optimization algorithm, MISCSO achieves a 9% improvement in overall performance. Comparative experiments were conducted between MISCSO and four recently improved algorithms. The results demonstrate that, in the trajectory optimization of the manipulator, the solution performance of MISCSO is superior to all comparison algorithms, and the standard deviation of solutions in multiple runs is the smallest, indicating optimal stability and robustness. MATLAB simulation results illustrate that the resulting joint motion curves are smooth, continuous, and without sudden changes.