Based on the analysis of the air strike made by a formation of UCAVs, the deployment optimization of the formation is discussed. Firstly, the attacking process is analyzed based on the real word scenario. Secondly, the Continuous Time Markov Chain theory is introduced to model the basic models and the probability transient models of the UCAV combat process. Thus the UCAV formation campaign result can be forecasted. The Multi-attribute method is used to integrate the data gained previously and the optimized deployment can be attained.