Aimed at the problem that the existing track clustering algorithm does not take into account the influence of aircraft heading change and altitude drop on the clustering results, at the same time, the clustering process lacks time information, and in addition, outlier data exist in the two radar data, and the existence of outliers will affect the final clustering result, leading to inaccurate clustering results, this paper proposes a time window segmentation algorithm based on feature point track. The algorithm is to take the aircraft heading change approach value and height decreased value as the influence factors of determining cluster size, the paper segments the number of aircraft entering the track point time window segmentation. The simulation analysis of real approach two radar data shows that when the influence factor is 0.4, the curvature of track is minimum, and the clustering effect is the best, then the hierarchical clustering algorithm is used to cluster the tracks corresponding to different values, and the final clustering results are obtained, providing technical guidance for the controller's scene command.