By leveraging the correspondence between spares configuration schemes and equipment support performance， the BP（back propagation） neural network model is used to adjust the mapping relationship between spares configuration schemes and their corresponding satisfaction and utilization rates. This facilitates the evaluation of the rationality of spares configuration scheme design. Using three types of equipment with different structural compositions as examples， the corresponding neural network prediction models for the satisfaction and utilization rates of spares are designed and trained with data samples. This achieves fast and high-precision calculations of the satisfaction and utilization rates of spare parts， with both the mean error and mean square error being less than 0.05%， demonstrating the effectiveness of the proposed method.