Abstract: The study addresses positioning delays caused by residual vibrations in high-speed SCARA robots through a vibration suppression method integrating geometric parameter optimization. A Lagrangian dynamic model incorporating key geometric parameters such as link length, joint spacing, and cross-sectional moment of inertia is established to reveal the intrinsic relationship between geometric parameter and vibration modes during high-speed motion. An optimization framework that minimizes positioning time with geometric constraints is developed and implemented using genetic algorithms to achieve dynamic mass balancing. Experimental results demonstrate a 32% improvement in efficiency, reducing positioning time from 4.804 s to 3.281 s while maintaining positioning accuracy. The proposed approach provides theoretical guidance for geometric design of high-speed SCARA robots, applicable to high-precision manufacturing scenarios such as semiconductor packaging and lithium battery electrode stacking.