Aiming at the inefficient and high-cost fresh commodity distribution, the work aims to study the fresh commodity distribution route optimization based on time windows and multi-compartment temperature control by adopting multi-compartment vehicles with temperature controlled compartments as distribution equipment and applying time windows and other constraints. Firstly, the bi-objective model was established to minimize logistics operating cost and the number of vehicles. Then, the non-dominated sorting genetic algorithm based on the Clarke-Wright saving algorithm (CW-NSGA-Ⅱ) was designed to solve the model. Among them, the initial population was generated by the Clarke-Wright saving algorithm, which improved the quality of the initial solution, and an elite iterative strategy was designed to improve the optimization performance.Based on the improved Solomon example, the proposed algorithm was compared with the multi-objective particle swarm algorithm, multi-target ant colony algorithm and multi-target genetic algorithm, verifying the solution performance of the CW-NSGA-Ⅱ. Combined with a case study, the indicators such as the number of multi-compartment vehicles, temperature control costs and operating costs were compared and analyzed. The results showed that the number of optimized multi-compartment vehicles was reduced by 35.7%, the temperature control cost was reduced by 39.2%, and the total operating cost was reduced by 47.7%. The proposed model and algorithm can effectively optimize the distribution route, reduce the total operating cost, and provide theoretical support and decision-making reference for the construction of the efficient and low-cost fresh distribution network.