Abstract: Abstract: Raman spectroscopy has been widely applied in some areas, such as agricultural products, food, and so on. In chemical molecular structure analysis and appraisal, it has some merits of simple pretreatment, nondestructive, rapid detection, etc. We all know that the Raman spectra of some substances with similar molecular structures may have significant differences. We can take advantage of this characteristic to effectively distinguish some substances with other similar molecular structures. Rice leaf blast is one of the most serious diseases that affect the yield of rice in cold area. In order to reduce the impact caused by the disease and increased early detection methods, Raman spectroscopy was used to collect spectroscopy of normal rice leaves and abnormal rice leaves and identified the characteristic frequency of the rice leaf. Through analysis of the spectral peaks and characteristic frequency offset of normal rice and abnormal rice functional groups, we found that the spectral peaks and the sensitive spectrum lines ranged from 1 800 to 2 600 cm-1. The slopes of bimodal lines between 984 and 994 cm-1, and between 828 and 851cm-1 were increased as the degree of disease increased. In random samples of 20 Raman spectra, the correct recognition ratio reached 75%. Collected Raman spectra have certain fluorescence background, and the process of baseline correction to the background has great influence on the Raman band peak height, but it has little influence on the peak area relatively. Therefore we can divide it by the peak maxima in the calculation of the spectral intensity. Finally, 50 samples of Raman spectra were randomly selected for analysis. We can get 1 005, 1 527cm-1 in the vicinity of the scattering cross section size and 2 000-2 300cm-1 within the scope of the multiple spectral sum of peak scattering cross section size. We can see the change from the first 25 samples which are not very large. But with the aggravation of the rice disease extent, the scattering cross section values increased obviously. It showed that the Raman spectra of the rice leaf without disease were very small in the 2 000-2 300 cm-1 range. Because in the acquisition process of the Raman spectrum of rice leaf, the total energy of the light source is conserved, the increase in rice blast disease sensitive band scattering intensity of spectral peak intensity will correspondingly weakened other characteristic functional groups. So, the influence of ratio of rice blast sensitive band scattering intensity and scattering peaks characteristic functional groups were more obvious in rice blast disease of Raman spectral lines. The scattering cross section of 2 000-2 300 cm-1 increased with the increase of the degree of disease through the analysis of 50 randomly selected samples of Raman spectra, which showed a good relationship between the changes of the scattering cross section and the rice blast disease. Raman spectroscopy provides an effective method for the early detection of rice leaf blast.