The doctoral thesis is dedicated to the development of the multi-agent system aimed at improving the efficiency of supply chain management. Supply chain management science is constantly evolving; new technologies are being introduced and new supply chain techniques and solutions are being offered. Although the application of multi-agent systems in supply chain management is not an innovation, it is still a topical research, especially in terms of re-use. The aim of the doctoral thesis is to develop an approach to the multi-agent system development that ensures the improvement of supply chain management efficiency. The developed approach allows creating a multi-agent system that can be used for various supply chain participants, as it reviews main supply chain processes and tasks to be solved. Various inventory management methods were used in the developed multi-agent system: forecasting algorithms, ABC analysis, and inventory replenishment point determination. Production scheduling and rescheduling were also used in the multi-agent system to reduce the total production time. During the development of the doctoral thesis, main efficiency indicators were analysed and indicators to test the efficiency of the developed multi-agent system were selected. Existing multi-agent systems for supply chain management and their reusability were investigated. It was concluded that there is no multi-agent system that covers all requirements and a multi-agent system needs to be developed that could be applied to different participants in the supply chain. As a result of the research on the supply chain tasks, it was proposed to create agents that cover main activities in each supply chain node, and the means of communication between nodes was chosen. The developed multi-agent system can be expanded with new agents or efficiency indicators if needed, or vice versa, agents that are not necessary can be removed from the multi-agent system. This multi-agent system was approbated on real company data and evaluated with efficiency indicators. The results of the practical application of the developed multi-agent system have proved its efficiency.