This thesis deals with design of a suitable approach for managing the supply of spare parts for machines approaching decommissioning. The appropriate balance between outstanding risks related to equipment downtime and costs of holding spare parts is predicted by proposed systematic algorithm. Determination of key parameters affecting the whole process of spare parts inventory management and their subsequent use as a basis for Event Tree Analysis lead to finding an optimum set-up, which mitigates the overall impact of the planned reduction in spare parts storage levels. The new approach has a number of advantages including versatility of application for many types of machines, as well as a different outlook on a solution for the current issue in spare parts management.