In MEMS LiDAR systems, external mechanical disturbances can lead to abnormal trajectory deformations and incoherent 2D point cloud sampling artifacts. To address these challenges, this study proposes a novel measurement framework that enables real-time suppression of trajectory errors through dynamic, adaptive adjustment of scanning parameters (amplitude, frequency, and phase) and uncertainty-aware estimation. Under external vibration conditions, the proposed method demonstrates superior stability compared to conventional approaches, achieving a maximum error reduction of 6.14%. Experimental results show that residual uncertainty is maintained within 5.51% (with a minimum error of 0.6%), the angular resolution of the scanning trajectory is preserved at high-precision level of θ=0.1497± 1.45%. The Root Mean Square Error (RMSE) of the correction algorithm is reduced by up to 5.46% relative to traditional methods, thereby fulfilling real-time processing requirements. Through theoretical modeling and validation on a mobile robot platform, the proposed approach exhibits strong anti-interference performance across varying levels of external vibration. The integration of dynamic adaptive calibration and uncertainty-aware estimation effectively addresses the core challenges in vibration-resilient LiDAR measurement systems, advancing applications in autonomous driving and intelligent robotic navigation.