In this paper, we consider a disaster scenario where multiple physical infrastructure components suffer damage. After the disaster, the health of these components continue to deteriorate over time, unless they are being repaired. Given this setting, we consider the problem of finding an optimal sequence to repair the different infrastructure components to maximize the number of components that are eventually returned to full health. We show that under certain conditions on the rates of improvement and deterioration, we can explicitly characterize the optimal repair policy as a function of the health states. We find that the optimal sequence depends on the relationship between the rate of improvement (when being repaired) and the rate of deterioration (when not being repaired).