According to the characteristics of multi-variety and freight loading problems with priority, establish target function fully utilizing capacity and cubage of loading tools. And solve this type of problems utilizing improved hybrid genetic algorithm from the overall situation. Firstly, adopt binary code so as to make the problem more succinctly. On the basis of cubage-weight balance algorithm, construct initial solution to improve the feasibility. Retain the best selection so as to guard the diversity of group. Secondly, the study adopts 2- exchange mutation operator, combine hill-climbing algorithm to strengthen the partial searching ability of chromosome. Finally, the example can be shown that the above model and algorithm is effective and they can provide for large-scale ideas to solve practical problems.