During disaster events, timely and targeted information provision and exchange could provide great help to the stricken population in difficult and complicated environments. This paper reports a serviceoriented system, called Rescue Wings, for providing emergency support to sufferers and rescuers in disasters. The system utilizes mobile services to acquire real-time information about the users and environment, and constructs service agents (servants) to provide active services for mobile users. To perform their functions, the servants frequently invoke a set of intelligent services of Rescue Wings, which can further access a number of public services from government and other public organizations. We identify the most frequent request sequence patterns (FRSP) of Rescue Wings, and develop a new bio-inspired algorithm for efficiently scheduling the requests to minimize the response delay. The system has been tested in several disaster rescue drills, and has been successfully applied to the 2013 Ya'an Earthquake in Southeast China.