To explore methods for evaluating the aging quality of tobacco lamina and establish a corresponding evaluation model, this study utilized middle tobacco leaves from the 2020 southwestern tobacco-growing region as experimental materials. Five aging time (10, 16, 22, 28, 34 months) were set up to determine 10 chemical composition indexes such as total alkaloid, total sugar, polyphenol, γ-aminobutyric acid, proline and neophytadiene. Analytical methods such as descriptive statistics, correlation analysis, principal component analysis (PCA), and regression analysis were employed to assess the data and construct an aging quality evaluation model. Results revealed variations in evaluation indices across different aging periods: proline exhibited the smallest coefficient of variation (CV) at 5.19%, while amide amino acids showed the largest CV at 25.65%. KMO (0.783) and Bartlett sphericity test (P<0.05) were suitable for principal component analysis. PCA extracted two principal components with a cumulative variance contribution rate of 82.30%. Finally, a regression model (R²=0.996) including 7 indexes of γ-aminobutyric acid, total alkaloid, polyphenol, proline, neophytadiene, total sugar and basic amino acid was established. The prediction results of the model were completely consistent with the sensory quality score. The model developed through PCA and stepwise regression analysis provides a theoretical foundation for evaluating tobacco lamina aging quality and determining optimal aging durations.