Abstract: This study aims to estimate the canopy chlorophyll density (CCD) of flue-cured tobacco with non destructive and accurate methods in real time to obtain the photosynthetic capacity and nutritional status. To this end, the relationship of canopy chlorophyll density and hyperspectral reflectance of flue-cured tobacco under different drought stresses was studied using an ASD spectrometer. Estimating models of the flue-cured tobacco canopy chlorophyll density were set up by means of the first derivative spectral reflectance. The results indicated that flue-cured tobacco canopy spectral reflectance showed orderly changes with chlorophyll density after drought stresses. The correlation between the first derivative spectral reflectance at 712 nm and chlorophyll density is the best (r=0.838). The BP neural network generated the best estimation. In the inversion of monadic linear model and BP neural network model for CCD using the first derivative spectral reflectance, the BP neural network model showed the best effect with the R² reached 0.9686 and RMSE being 0.0778. The results may provide the basis for the cultivation and management through long-term real-time monitoring of the photosynthetic capacity and water stress status of the flue-cured tobacco flourishing population.