Forty six points representing different severity degree of stripe rust were established in winter wheat field. The canopy reflectance was collected by an ASD hand-held spectrometer at each point. Meanwhile,the diseases index was investigated. These data were used for the following analysis. Firstly,the relationships between diseases index and reflectance of bands in the range of 300-1 500 nm were analyzed. The sensitive bands were selected for stripe rust detecting. Secondly,considering the character of PHI image,red bands (620-718 nm) and near infrared bands (770-805 nm) were assigned as the best bands. Finally,the mean reflectance of red bands (620-718 nm) and near infrared bands (770-805 nm) was calculated respectively to construct the reverse model with the observed diseases indexes,DI=19.241R1-2.206 67 R2+12.274 4. With this model,the severity degree of stripe rust of winter wheat was monitored successfully in PHI image.