At present, the civilian aircraft cargo hold is mainly used photoelectric smoke detector as a fire smoke identification, due to the aircraft cargo hold will be suspended in other interference particles (dust, or water vapor, etc.) will trigger the fire detector alarm, there is a high false alarm rate problem. Based on the dual-wavelength photoelectric smoke detector can effectively reduce the problem of high false alarm rate, the collection of aircraft cargo hold typical combustibles as well as interference sources of infrared light scattering power, blue light scattering power as a fire parameter, through the particulate matter of infrared light scattering power and blue light scattering power analysis, as well as a collection of the Sauter average particle size analysis for the subsequent development of composite smoke detector optimization of the fire parameter, the results show that: the infrared and blue light have a good effect on the fire smoke particles, the results of which are: infrared and blue light have a good effect on the fire smoke particles. blue light have a better response to fire smoke particles, combined with the average particle size of Sauter can effectively distinguish between fire smoke particles and non-fire smoke particles, reduce the false alarm rate, the false alarm rate of the interference source does not exceed 6.7%.