Since a vehicle-borne light detection and ranging (LiDAR) measurement system is affected by the signal shielding and rolling vibration of the vehicle as it moves, commonly available trajectory data are usually low-quality data with noise. Although the position and attitude data of the trajectory are processed by joint Kalman filtering, there are still fluctuations in local areas, which require processing to smooth the acquired trajectory data. In this paper, a model of vehicle motion is proposed to analyze the trend of a vehicle trajectory over time by recording the vehicle position, velocity, and attitude information in real time. Next, the vehicle motion trajectory is processed in sections to understand the motion state intuitively. Experimental results show that the descriptions of ground features by the vehicle-mounted and ground point clouds are almost the same. The relative accuracy can reach 0.013 m by selecting multiple spacings for comparison. In conclusion, the segmentation optimization method for vehicle trajectories proposed in this study is expected to provide a higher accuracy than the current techniques used for optimizing the accuracy of vehicle trajectories.