Thousands of victims and millions of affected people are consequences of natural disasters, every year. Therefore, it is essential to prepare a proper response program that considers early activities of disaster management. In urban areas, local organizations and municipal authorities are responsible for disaster-related operations such as providing emergency shelters, proper equipment, and relief supplies in order to reduce sufferings of survivors. In this paper, a new multi-objective mixed-integer linear programming model is developed for preparation planning phase of disaster management. The proposed model is inspired from a real case study of an urban district in Iran, which considers both humanitarian- and cost-based objectives in a goal-programming approach. Proactive damage estimation result of Risk Assessment tool for Diagnosis of Urban Areas against Seismic Disaster software is used as an input in order to allocate affected people to local emergency management facilities that should be established. The location allocation model is solved for both current municipal subregional zoning and a virtual zoning approach that creates auxiliary cells. Mathematical results show that the second approach can reduce logistic costs and increase total coverage simultaneously. In other words, using virtual zones would help the authorities create a better collaboration between neighboring local areas and as a result, efficiency of the decisions improves.