A multi-spectral earth observation LiDAR is a new means to monitor vegetation, and its applications are increasingly widespread. The existing multi-spectral earth observation LiDAR system adopts blazed grating and multi-channel photosensitive detector arrays as its data reception mode. Increasing the number of receiving channels only to improve the spectral resolution and receiving range will cause such problems as data redundancy, difficulty in system integration and high cost etc. How to determine the number of channels and how to adjust the characteristic wavelength just to the center of the appropriate channel by selecting blazed grating and its blazed wavelength is discussed; in the meantime the characteristic wavelength outside the channel or not at the center of the channel is amended by using the feature weighting method based on the principal component analysis. Thus, not only the effectiveness on receiving information is improved, but also the wavelength selection theory is supplemented to some extent while spectral information of vegetation is not reduced, and the adaptation of the multi-spectral earth observation LiDAR system in practical applications is improved as well.