Abstract: In order to realize fast and real-time monitoring of stomatal conductance of flue-cured tobacco leaves with spectral information, two varieties, K326 and Zhongyan 100, were used as experimental materials to analyze the relationships between the stomatal conductance of leaves and the original spectra or the first derivative of the canopy spectra under the conditions of regular water needed, mild drought, moderate drought and severe drought. The best sensitive bands for the stomatal conductance of the flue-cured tobacco samples were screened, and the estimation model for the stomatal conductance was established. The results showed that: 1) With the increase of drought stress, the stomatal conductance of flue-cured tobacco leaves and the spectral reflectance of the canopy at 700- 1 350 nm presented a decreasing trend, and the stomatal conductance at the fast growing stage was higher than that at the root extending and maturity stages, and the two varieties had the same performance pattern. 2) The sensitive bands were mainly within the near-infrared region. The correlation coefficient between the original spectra of the canopy and the stomatal conductance of the tested leaves at 715-930 nm and 963-1 000 nm were more than 0.6, and the correlation coefficient between the first derivative of the spectra at 712-715 nm and 718-719 nm and the stomatal conductance of the leaves was more than 0.7, all the correlations reached an extremely significant level. 3) The BP neural network model established with the first derivative of the spectra at 711-715 nm was effective, and the determination coefficient R² of the model was 0.908. Therefore, the first derivative of the flue-cured tobacco canopy spectra could reflect the stomatal conductance of the tested leaves more accurately, and its modeling effect was better than that of the original canopy spectra, in which the stomatal conductance model of the leaves established with the first derivative of the spectra at 711-715 nm showed the best effect.