In order to ensure the smoothness and safety of cargo in transportation and optimize the problem of container loading layout in air transportation, an improved genetic algorithm is proposed and the research of air container loading application is carried out. Considering seven practical constraints of cargo loading, taking the container volume utilization rate as the optimization objective, the multi-container loading optimization model of air container is established. The initial population is randomly generated by three-stage real coding, and the optimal individual protection strategy is added to enhance the global convergence of the genetic algorithm. A reasonable fitness function is constructed by combining different constraints. Taking the real air cargo information as experimental data, the experimental results show that the average utilization rate of container volume is increased from 74.07% to 83.99%, and the number of loading pieces is significantly increased, which is suitable for air container transportation. The algorithm can be applied to air container loading and transportation, creating conditions for the realization of intelligent loading and the improvement of transportation efficiency in the air transportation industry.