It is difficult to determine the health of rice for simple vegetation index (NDVI) thresholding method which is widely used through remote sensing technology in crop disaster monitoring. The study selected binary logistic regression which were respectively established vegetation index getting from measured spectral and the relationship model between health status. The results show that the triangular vegetation index TVI model is with better reliability. When remote sensing monitoring rice blast was taken into account, geographical range was widely involved and rice-growing conditions were existing obvious differences in the local area, using a 3 × 3 pixel neighborhood consistency assumption to eliminate differences in the local environment. Applying China’s own property "environmental disaster satellite" CCD sensor data into the model and the stress range of extracting rice blast was basically consistent with Plant Protection Institute of Heilongjiang Academy of Agricultural Reclamation Sciences as well as the ground measured results, among which TVI model results accuracy reached 76.47%, which can meet remote sensing monitoring requirements. of the blast area.