This paper develops and analyses two basic models for hospital location and capacity allocation. The focus is on an area prone to natural disasters. The first model seeks to locate hospitals and allocate capacities so that the mean travel distance for patients to hospitals is minimised over a variety of disaster scenarios. The second model seeks to reallocate capacity among hospitals so as to maximise the system’s effectiveness to the upcoming disaster event. Heuristic solution methods of the two models are investigated, so as to make the approach computationally viable and to gain insight into the location and capacity allocation strategies. A regional planner can use these models in various ways. For earthquake prone areas (where there is little forewarning), the first model’s results can be compared to the current hospital locations and capacity allocations. A plan can then be developed to shift capacity between hospitals or in an extreme case to relocate hospitals so as to be better prepared for a disaster event. For hurricane prone areas (where there is considerable forewarning), the second model can be used to develop a plan for reallocation of capacities between hospitals in anticipation of the event. Furthermore, the first model can be used to better locate hospitals and select capacities for an area that is being rebuilt after a large magnitude hurricane has largely destroyed the area (as Hurricane KATRINA did to New Orleans in 2005). The results are illustrated with the help of case studies – one based on earthquake scenario in Northridge, California, and another based on a hurricane scenario in New Orleans.