This research presents significant improvements to the hospital capacity planning models for hurricane disaster proposed by Paul and Batta. Firstly, we develop a model that reallocates hospital capacities taking into account the effect of casualty severity on medical care demands and travel time to hospitals on casualty outcomes. Secondly, travel time to hospitals, a key input to the model, is estimated using a stochastic simulation. Thirdly, in order to meet with challenges that arise as disaster evolves, we develop an algorithm that dynamically updates the pre-disaster plans to post-disaster reality. The algorithm evaluates prior decisions by dynamically updating the hospital status, travel times and status of the casualty clusters. We demonstrate the application of models in developing pre-disaster plans (hospital capacity and ambulance reallocation) through a case study on New Orleans. In addition, we show the application of the algorithm for post-disaster planning via illustrative examples.