•The study proposes an optimized time window constraint model, incorporating multi-punishment factors, to balance multiple factors in path planning for cold chain transportation of perishable agricultural products.•The improved Ant Colony Optimization (IACO) method is used to generate distribution paths with shorter distances, providing new technical support for perishable agricultural product cold chain distribution path optimization.•The proposed model shows better computational efficiency compared to traditional Genetic Algorithm (GA) methods. With the development of modern logistics industry, the cold chain logistics distribution of perishable agricultural products has become an important link to ensure food quality and safety. Aiming at the complicated situation in the process of perishable agricultural products transportation path planning, a cold chain logistics distribution path optimization model based on Ant Colony Optimization (ACO) was proposed. Based on the traditional soft and hard time window constraint model, an optimized time window constraint model was proposed, and the update function of pheromone increment was substituted to optimize the traditional ant colony algorithm. An Improved Ant Colony Optimization (IACO)-time-window constrained path optimization model was proposed. The test results showed that in the convergence training experiment in Matlab environment, the research method only needed 19 iterations to reach the lowest level. In the experiment of application effect, the length of the path generated by the research method for car a was 188.93 km. Car Group a could complete 60.38 % offload at the fourth distribution node. The experimental results showed that the proposed route optimization model had better iteration efficiency and could effectively generate a shorter distance distribution path, which provides a new technical support for the route optimization of cold chain distribution of perishable agricultural products. In this paper, a path optimization model of cold chain logistics distribution was proposed. By improving the time window constraint and introducing the update function of pheromone increment to optimize the ant colony algorithm, the path planning optimization problem in the cold chain logistics distribution of perishable agricultural products was solved. The results show that in the total cost convergence test, the initial total cost function value of the research method is lower than 3245. In the practical application analysis, the total path length of the two transport vehicles is 188.93 km and 214.44 km respectively. Display omitted