The work aims to meet the requirements of high freshness when fresh products are delivered, and realize the optimization of location routing of distribution center under the scenario of multiple products and multiple vehicle models. A two-level goal programming model was constructed in view of considering the cost of carbon emission and meeting the customers' requirements for the minimum freshness of product at delivery. The optimization goal of the upper model was to minimize the fixed cost of distribution center and inventory management cost, while the optimization goal of the lower model was to minimize the fixed cost of vehicles, transportation cost, carbon emission cost and punishment cost. Combined with the characteristics of the model, a two-stage heuristic algorithm was used to solve the problem. The average cost improvement rate of the two-stage heuristic algorithm was 1.22% compared with that of genetic algorithm, and 3.03% compared with that of K-means clustering algorithm. The average improvement rate of the two-stage heuristic algorithm was 24.8% compared with that of genetic algorithm, and 33.0% compared with that of traditional K-means clustering algorithm. Through the comparative study of calculation examples, different freshness requirements have a significant impact on the location of distribution centers and the arrangement of vehicle routes. Enterprises can reduce logistics costs by means of reasonable logistics network planning and accurate evaluation of customers' requirements for product freshness.