The fleet anti-air defense system is growing increasingly complex. This complexity requires a complicated doctrine for its operators, and the doctrine needs to be analyzed under what-if scenarios. The most ideal analysis method is investigating live combat outcomes, yet such events rarely occur. Hence, we develop a discrete event-based model, perform battle experiments of fleet anti-air defense, and analyze the results. This battle experiment resulted in a better fleet formation against air threats and recommendations for weapon loads on warships.