Emergency Medical Services (EMS) attracted many researchers because the demand of EMS was increasing over time. One of the major concerns of EMS is the response time and ambulance despatching is one of the vital factors which affects the response time. This paper focuses on the problem of ambulance despatching when many emergency calls emerge in a short time, which exists under the condition of catastrophic natural or manmade disasters. We modify a new method for ambulance despatching by centrality measure, this method constructs a nearest-neighbor coupled emergency call network and then prioritize those calls by the score of fitness, where the score of fitness considers two factors: centralized measure a call by the emergency call network and the closest policy which means despatching to the closest call site. This method is testified by a series of simulation experiments on the real topology road network of Hong Kong Island which contains 8 hospitals. These analyses demonstrate the real situation and proof the potential of centrality measure in reducing response time of EMS.