The work aims to constructa joint optimization model for vehicle loading and distribution routes with the shortest route and the highest loading rate, taking into account constraints such as the indivisibility of orders and the volume of goods, so as to address the delivery route problem with common limitations of quality and volume. In terms of solving the vehicle route optimization model, first the clustering algorithm was used to partition the distribution area, and then the load capacity of the vehicle was used to determine whether the station goods could be delivered. Finally, the genetic algorithm was used to obtain the optimal route. The greedy algorithm and the block based heuristic algorithm were used to solve the loading problem of goods in the 3D loading model. The feasibility of the model and the algorithm was verified based on a specific example of a company. After optimization, the number of vehicles for delivery was reduced by 1, the delivery distance was reduced by 154.247 km, and the average loading rate reached 93.89%, saving the company's delivery costs. The results indicate that the constructed model and the solved algorithm can improve loading rate and delivery efficiency, providing a theoretical basis for solving the joint optimization problem of vehicle loading and distribution routes.