Abstract: With the rapid development of intelligent transportation, how to generate high-precision maps based on vehicle trajectory data has become a major problem in the industry. Trajectory similarity calculation and similar trajectory extraction are important steps to use the similar trajectory to infer the remaining lane position, and its correction greatly affects the accuracy of the lane position. The traditional LCSS (longest common subsequence) algorithm is mostly used to calculate the similarity of overlapping trajectories. To solve this problem, according to the characteristics that the lane trajectories are parallel and maintain a fixed distance, an improved LCSS method is proposed. Firstly, the buffer is constructed to screen out similar trajectories, then the trajectory alignment strategy based on translation and resampling is used to synchronize the two trajectories in space and time. Finally, the similarity of the two trajectories is calculated based on LCSS. When the similarity satisfies the threshold condition, it is determined that the trajectory pairs are similar. Experiment results show that the proposed method can effectively extract similar trajectories.