The missile boost section has high temperature and large area of plume. The missile can be accurately identified according to the spectrum characteristics of the plume. Applying the plume spectrum into the space-based early warning system can improve the missile defense ability, but the huge spectral information results in a lot of time from obtaining spectral information to identifying spectral information. So, we find the characteristic spectrum which can represent the spectrum of the missile plume, and use fuzzy algorithm to process the characteristic spectrum data, so as to achieve the purpose of accurate and rapid identification. Analysis shows that the characteristic spectrum image has higher signal-to-noise ratio than full band infrared image, and fuzzy algorithm is better than spectral angle measurement (SAM) on real-time and accuracy.