The container loading plan significantly affects the transportation cost of manufacturing companies. However, the traditional approaches for three‐dimensional single container loading problems cannot be directly applied to three‐dimensional multiple containers loading problems (3D‐MCLP) due to the different basic features between the two problems; in addition, the existing solution algorithms for a 3D‐MCLP have limited efficiency and quality when dealing with nonidentical boxes corresponding to multiple customers. Considering a general 3D‐MCLP with weak heterogeneous containers, this paper establishes a general mathematical model and proposes an integrated greedy algorithm, multilayer tree search, and large neighborhood search algorithm to maximize the average filling rate of containers. An integrated greedy heuristic multilayer tree search algorithm is proposed to obtain the initial solution of a 3D‐MCLP. Then, a large neighborhood search algorithm containing five removal operators and two repair operators is further designed to improve the results. Numerical experiments with 2014/2015 ESICUP challenge datasets and real‐world case studies validate the effectiveness and efficiency of the proposed approach.