Aiming at the three-dimensional container loading problem with maximum space utilization as the goal, this research proposes an improved genetic algorithm based on the priority retention strategy to solve it. First, based on the analysis of existing related research deficiencies, the basic improvement concept of the priority retention strategy is proposed. Second, the basic process of improving the genetic algorithm is designed according to the characteristics of the problem, focusing on the detailed implementation of crossover and mutation. Finally, through experimental simulation, the solution results and algorithm performance are compared and analyzed. Experiments have proved that the algorithm has better solution results (about 14% better than the compared algorithm), the convergence is more stable, and energy consumption meet general needs. Therefore, this algorithm has relatively excellent iterative performance.