During the research, the authors developed an advanced solution for determining the optimal size of reserves, as well as a reserve management procedure and algorithm. The developed solution considers such parameters as stock circulation, classification, stock balance, order point, guarantee stock, stock order and filling time, allowable delay in order execution, and maximum stock value. Mathematical modeling methods were used to predict the demand for stocks in the warehouse. The developed model and algorithm determine the management decision about “when and how much stock to order.” According to the variable demand, the results of the stock management algorithm are obtained, and its performance is presented. The use of the stock management algorithm made it possible to reduce the input costs of goods of the same nomenclature in the warehouse of the wagon depot by 3% and the volume of residual stocks in the warehouse by 60%. Manufacturing companies that overorder goods end up with unnecessary costs.