This chapter goes one step further from spatiotemporal points to their connected components—trajectory. Learning and analyzing dynamic trajectories can characterize the behavior of various moving targets in a scene and provide a grasp of scene changes and motion states. Trajectory contains not only the spatial location of spatial points, but also the velocity, acceleration, etc. of spatial points. Trajectory considers not only the global connections among points, but also the local features and relation among points. With so many information available, the work of spatiotemporal trajectory analysis includes trajectory detection, tracking, modeling, learning, classification, recognition, etc. This chapter will introduce the basic principles and steps for trajectory analysis, emphasize modeling and learning. Some trajectory related activity analysis examples are also provided. This chapter will also introduce a method for extracting trajectory feature, classifying and recognizing trajectories of different target movements by using trajectory feature clustering trees. Experiments have been made on several databases with several comparable methods, and the results obtained show the effectiveness of this method with the evaluation indicator—mean Average Precision (mAP).