Abstract: In order to achieve accurate, fast and efficient grading of tobacco by a more scientific and systematic way, reducing sugar/total nitrogen ratio (Kovalenko index), total nitrogen/nicotine ratio and nicotine mass fraction were used as the main evaluation indexes for tobacco quality. An intrinsic quality evaluation model for Yunnan tobacco leaves was developed based on the Chi-square function, which is non-linear, non-monotonic and asymmetric, and the distribution of chemical indexes of the tobacco leaves in redried modules with different appearance and sensory quality grades. Based on the developed model and the near-infrared spectral data, the intrinsic quality of leaves in different redried modules of Hongta Group was evaluated. The results showed that: 1) The results were highly related to the trends of tobacco appearance and sensory quality classification. 2) The results were improved to different degrees after blending the leaves of lower quality scores with the leaves of complementary chemical indexes in certain proportions (e.g. the upper leaves with higher nicotine mass fraction and the lower leaves with lower nicotine mass fraction). Among them, the combination of leaves of BO and XL grades could promote the intrinsic quality of the blends to the greatest extent. 3) The developed model could achieve rapid and efficient grading of tobacco intrinsic quality by a few key chemical indexes. Furthermore, the application of the intrinsic quality evaluation results of industrially graded tobacco could provide guidance for the formulation of the redried modules.