At the current stage in the development of domestic industrial enterprises, it is especially important to ensure reliable operation of the corresponding supply networks. Review and comparative analysis of literature in the respective scientific area allows to conclude that relatively few methodological procedures and instrumental tools have been for determining the performance characteristics of supply networks, including their reliability parameters. It was thus deemed practical to create an optimization model for determining these characteristics for the supply networks of industrial enterprises based on a generalized setting for the problem of managing a supply network. The model created allows to determine the network’s optimal configuration with respect to the composition of the participating enterprises and the material flows processed by these enterprises on the basis of the criteria describing the general costs of implementing the project (as ordered by the ultimate consumers), the total duration and the generalized reliability of implementing the respective processes concerned with product transfer and processing. Each of the above-mentioned criteria can be taken into account within the model either as a component of the objective function, or as an element of the corresponding constraints. The main initial data for implementing the optimization model are the characteristics of the supply network’s process model concerning costs, duration and reliability characteristics for product transfer and processing, and also the demands of the ultimate consumers for the amount of the finished product, general costs, total duration and reliability of fulfilling the order (the last three characteristics can be determined by both normative and maximum admissible values depending on the method used for taking into account the optimization criteria). The main sought-for variables of the model are the volumes of material flows between the enterprises making up the supply network; the auxiliary variables describe cumulative values of duration and reliability at different processing stages. As opposed to most of the previously developed tools, this optimization model is linear and therefore can be easily implemented with modern software tools.