One of the uses of reinforced concrete pipes (RCPs) is the distribution of aggressive water in industrial systems, for example, in water-cooling systems of nuclear power plants. Some of them carry seawater and can deteriorate with time because of internal corrosion. Because of the low O(2) content of aggressive water, slow corrosion is expected for such applications. If the RCPs are not periodically replaced, they will eventually fail. Replacement strategies for these pipes depend on (1) the risks associated with the failure of the water distribution network, and (2) the costs associated with replacing the pipes, including the removal of existing pipes, installation of new pipes, and associated production losses. Because of the lack of statistical data regarding RCP failure, the development of a risk-based replacement strategy is not an easy task. This article demonstrates how predictive models for the evolution of the failure of RCPs and the associated consequences of failure can be used to develop risk-based replacement strategies for RCPs. An application for the replacement strategies of a network modeled as a system consisting of 228 RCPs is presented as a case study. We focus on the assessment of the number of replaced components that governs the costs. The main objective of this article is to provide a theoretical approach for comparing replacement strategies, based on (1) the results of a reliability study, (2) the representation of the distributions of failed components (binomial distribution), and (3) the decision tree representation for replacement of RCPs. A focus on the scatter of the induced costs themselves is suggested to emphasize the financial risk.