The works aims to improve the utilization rate of packing cases in the packaging link of e-commerce warehousing. Aiming at the 3D packing problem of multi box and multi-item types in e-commerce warehousing, a mathematical model of mixed integer programming was established, and a packing process module algorithm based on heuristic empirical rules and the combination of multiple operators was designed. The hybrid genetic algorithm for multi box 3D packing problem was designed from packing sequence and improved operator to optimize the packing scheme. Experiments showed that PSO-HGA-S1 algorithm was the best among the PSO-HGA algorithm series in the optimization of packing sequence. Among the hybrid genetic algorithms with improved operators, IPO-HGA-S1 algorithm was the best. The hybrid genetic algorithm designed in the work can improve the utilization rate of packing cases in the packaging link of e-commerce warehousing.