In order to overcome the short comings of the fresh good logistics distribution optimization study in the reasonable combination of customers’ demand time windows and fresh good temperature control, and considering the characteristics of high timeliness of fresh goods, this paper establishes a bi-objective optimization model including the minimum logistics cost of fresh good distribution and the minimum loss of fresh goods value. Firstly, the logistics cost model containing transportation cost, temperature control cost and penalty cost of the time window is established, and the fresh value loss model based on temperature control is established. Then, according to the characteristics of the model, the K-means clustering algorithm is designed to consider the customer space location, the demand commodity temperature control and the time window constraint, therefore a genetic algorithm-tabu search(GA-TS) hybrid algorithm is proposed. This hybrid algorithm combines the global search capability of GA and the local search capability of TS. Through the comparison with the hybrid genetic algorithm, the multi-objective particle swarm optimization algorithm and the improved ant colony optimization algorithm, the validity of the model and the algorithm is verified. Finally, the best distribution temperatures of the fresh good and the best number of clustering schemes are obtained through sensitivity analysis. The results show that the model and the algorithm are reasonable and effective, which can provide reference and decision support for the logistics enterprise's fresh good distribution optimization.