Abstract: Residual vibration induced by joint flexibility during high-speed operations of industrial robots significantly constrains production efficiency. While traditional input shaping techniques are effective for vibration suppression, their inherent convolution operations introduce fixed time delays and result in excessive computational complexity for the inverse calculation of shaped trajectories. To address these issues, an improved Double S trajectory planning algorithm that incorporates frequency-domain shaping is proposed. The method establishes a mapping between trajectory parameters and spectral zeros, thereby transforming vibration suppression into a parameter optimization problem. By deriving seven critical displacement conditions, an adaptive strategy comprising eight modes is constructed. This approach utilizes the impulse cancellation principle to overcome time constraints in short-distance vibration suppression while ensuring full-range time continuity. Experimental results on an ER35-1800 robot demonstrate that, compared with standard planning, the proposed method reduces the maximum position settling time by 45.8% and the maximum position overshoot by 64%.