Health care centers are of the most urban users which are directly involved in supplying individual and social health. Fast, inexpensive and right on time access to these centers is of significant importance in any society particularly in urban society. The main duty of programmers and decision makers in determining urban optimal location of these centers is in the way that all urban population can have access to these centers easily. In this research location problem of hierarchical health care facilities is considered. Due to indefinite nature of demand, a mathematical model relied on fuzzy credibility theory which represented the amount of demand as a triangular fuzzy variable. Since Iran is among ten disaster-prone countries and sixth seismic country, though it is impossible to compensate harms from different perspectives but it is possible to minimize harms by precautionary actions and suitable programming in confronting of events. For instance one the important strategies in order to improve system’s operation and to decrease delay in relief is location of health care centers in relief supply chain. Strategic importance of this discussion from different social and economic perspectives expresses necessities of paying attention to this problem. In this research in facing with uncertainty in critical situation an optimization model depending upon mulvey model is offered. This model includes determining health care and relief centers at different levels and by different services. Furthermore in order to help to injured people and to avoid congestion in hospitals and medical centers, there must be possibility of establishing temporary facilities and outpatients treatment centers to give initial emergency services also sending medical teams and equipment to these centers. Since the considering problem is among NP-complete problems cluster, harmony search algorithm and taboo search algorithm(by using two strategies in a way that in the first strategy this algorithm is mixed with variable neighborhood search algorithm) are used to give solution. Taboo search algorithm in two strategies uses a simple mathematical model to initiate the initial solution. In order to improve performance of the algorithms in optimization problem, Taguchi method is used. The lower bound of Langrangean relaxation method is used to compare meta heuristic algorithms for large dimensional problems. In following some sample problems in small and large dimensions accidentally are generated and results of exact solution and offered meta-heuristic algorithms, considered and compared together. Results show that taboo search algorithm (the first strategy) has better efficiency than harmony algorithm.