The recognition of flight maneuvering is primarily utilized for assessing pilot training quality， aiding decision-making in aerial combat， and other related scenarios. This study focuses on realizing flight maneuver recognition through the utilization of the bag of patterns （BoP） algorithm， which has been enhanced to address deficiencies in multi-dimensional time series scenarios. The improved algorithm extracts features from flight attitude data and analyze flight maneuver recognition. Simulation results demonstrate that the enhanced BoP algorithm enhances the accuracy and confidence of flight maneuver recognition， while effectively characterizing specific flight maneuvers through extracted parameter features.