The article concerns the improvement of the ACO (Ant Colony Optimization) ant colony optimization algorithm for the formationof routes of vehicles for the procurement of food raw materials on the territory of the community during emergencies. The purpose of the study is to improve the algorithm for the formation of routes of vehicles for the procurement of food raw materials on the territory of the community during emergencies. The proposed algorithm is based on the classical algorithm of ant colony optimization ACO and, unlike it, takes into account real production conditions during emergencies. The task of the research is to create an algorithm for the formation of effective routes of vehicles for the procurement of food raw materials in the territory of the community during emergencies, as well as its comparison with the classic ACO algorithm for solving various problems of route formation. It was established that the use of the classic algorithm for the optimization of ant colonies ACO, or its known modernizations, does not provide a high-quality solution to the problem of forming routes of vehicles for harvesting food raw materials on the territory of the community during emergencies.This is due to incomplete consideration of specific production conditions. The improved route formation algorithm involves 8 steps and is based on the classic ACO algorithm. In contrast to it, it takes into account real production conditions (damaged sections of the roadway,the presence of partial passage of vehicles, traffic jams caused by an emergency, etc.). The rule of the classic ACO algorithm regarding the selection of the next point in the route using the probabilistic-proportional transition of the k-th ant from the i-th to the j-th node (farm producing food raw materials) is proposed, replaced by one that takes into account the state of production conditions (road surface) be-tween individual nodes. This ensures an increase in accuracy and a decrease in the duration of route formation, as well as an increase in the quality of making appropriate management decisions. The obtained results regarding the comparison of the use of algorithms when solving transport problems with a different number of vertices indicate that the proposed algorithm provides a deviation of the total path in the route, which does not exceed 1%. The proposed algorithm reduces the decision-making time by up to 6% in the presence of up to 50 units of vertices, and by 12...15% in the presence of vertices from 51 to 100 units. The improved vehicle routing algorithm can be used in decision-making support systems to plan the procurement of food raw materials on the territory of the community during emer-gencies, which will increase their efficiency.