In order to optimize the allocation of flight slot resources and improve the regularity of civil aviation flights， the method of season division under the level of large-scale air traffic operation is studied in this paper. The monthly time series similarity of air traffic operation efficiency index is analyzed by the dynamic time warping algorithm， an optimized flight season division model is established， and the fast non-dominant sorting genetic algorithm with elite strategy is used to solve the problem. The TOPSIS method with entropy is used to evaluate the Pareto solution， and the final flight season division results under different division angles are given. Then and the effects of season division results and weight setting on the results under different air traffic operation levels are analyzed. The results show that， in the case of comprehensively considering the influence of the evaluation index of air traffic operation level， the result of optimal flight season division is stable with the change of weight. The method proposed in this paper is not only suitable for the flight season division of large-scale airspace environment， but also suitable for exploring the flight season mode of individual airport operation， thus providing a means for the Civil Aviation Administration and the airport to make flight schedule scientifically.