Maintenance of civil infrastructure is necessary to ensure assets perform as intended over their service life. As these elements age and degrade, appropriate repair planning helps ensure continued operation. Unfortunately, optimal repair planning can be complex due to uncertainty regarding component conditions. This uncertainty also makes budgeting difficult. This paper presents a novel methodology for optimizing maintenance planning of facility components under uncertainty. The proposed methodology integrates a Markov chain Monte Carlo process to capture the uncertainty inherent in equipment degradation, and a mixed integer programming model to generate an optimal repair plan and an associated minimized budget. This approach identifies an optimal set of facility components that need to be repaired to minimize the total facility repair cost required to meet or exceed a user-defined goal for facility health. The proposed method has distinct advantages over many non-linear approaches in literature as it can guarantee an optimal solution with significantly lower computational time. This computational advantage enables the technique to scale to more significant problems, such as determining repair plans across a portfolio of facilities. The method outperforms basic decision heuristics used to define repair plans and existing non-linear optimization methods in terms of consistently meeting a facility health metric at a minimal cost. Data obtained from the U.S. Army Corps of Engineers is used within a case study to demonstrate the feasibility of the proposed approach as it applies to building infrastructure maintenance.