Natural disasters (i.e. earthquakes, floods, hurricanes, lightning, avalanches, tornadoes, fires and volcanoes) and unnatural disasters (i.e. wars, terrorist incidents, road traffic accidents, industrial accidents, political unrest and refugees migration), are one of the main obstacles for sustainable development of countries. Therefore, unpreparedness and inappropriate encountering with them causes human causalities and heavy damages which is sometimes irreparable. In this paper, a mathematical model is developed for planning emergency logistics problem. The model aims in making decisions such as determining the location of distribution and relief centers and dynamic routing of aerial vehicles in order to supply the required goods of affected people in disaster. With regard to the situation of affected areas such as dire need to the relief goods and the importance of servicing time, aftershocks, inaccurate estimation of damages and casualties, routes network damages, new affected areas can be added to the problem during routing and demand of areas is considered dynamic. As reducing the service time in disaster plays an important role in reducing deaths and disabilities, therefore the proposed mathematical model aims in minimizing the maximum traortation time of relief goods to the located facilities. Since, covering tour is an approach which helps in increasing the servicing speed in disaster logistic systems and specifically in vehicle routing of necessary goods, thus it is used for vehicle routing in this problem. By considering covering tour approach, vehicles visit a limited number of nodes and outpatient relief and distribution centers are established in cross docks. These centers, in addition to the servicing of the injured people of that area, service to areas that are located near them. Also in this model, aerial vehicles are used in order to send relief goods. According to NP-hard complexity of proposed model, methheuristic algorithm such as genetic, basic scatter search and a modified hybrid algorithm based on scatter search and variable neighborhood search algorithm is developed. To study the efficiency of proposed algorithm, results obtained from exact method and methheuristics were compared. Results have shown that the proposed hybrid algorithm has a better performance in comparison with other algorithms.