Integrated Sensing and Communications (ISAC) has emerged as one of the key technologies for 6G communication networks. However, achieving high-precision sensing is hindered by multipath reflections in urban environments. Traditional methods require iterative estimation and separation of multipath signals, leading to high computational cost and poor scalability as the number of reflectors increases. To address this issue, we propose a Rapid Perturbation-based Multipath Error Cancellation (RP-MEC) method. The core idea is to superimpose a rapid perturbation at the receiver, inducing high-frequency oscillations in the multipath error for self-cancellation. Furthermore, this paper provides analytical expressions for the multipath error of Orthogonal Frequency-Division Multiplexing (OFDM) signals both before and after uniform perturbation, based on which the minimum required perturbation amplitude is determined. Experimental results validate the theoretical analysis and show that millimeter-to-centimeter-scale perturbations, whose physical scale depends on waveform parameters and path geometry, can reduce the multipath error envelope by 98.7% while maintaining sub-second update rates, thereby meeting the positioning requirements of mobile ISAC users. Compared with conventional iterative techniques, the proposed RP-MEC method not only significantly reduces computational complexity but also substantially enhances environmental robustness, making it a viable solution for high-accuracy sensing and localization in future 6G ISAC systems.