In order to provide a scientific basis for homogenized tobacco leaf production and soil fertility improvement for Xuanwei tobacco-growing area, 734 representative tobacco fields from 21 townships of Xuanwei tobacco-growing area were selected as research objects. Soil pH, organic matter, alkali-hydrolyzable nitrogen, available phosphorus, and available potassium contents were used as evaluation indices for soil fertility quality. A fuzzy membership function model was applied to construct a comprehensive soil fertility index (IFI) for quantitative assessment of soil fertility quality in the study area. Additionally, the correlation between the coefficient of variation (CV) of each soil fertility index and CV of IFI was analyzed. The results showed that, the overall soil fertility quality in Xuanwei tobacco-growing area was at a relatively optimal level, but significant differences existed among different townships. Regarding individual fertility indices, CV values varied notably: available phosphorus had the highest CV, followed by available potassium, while soil pH had the lowest CV. Correlations were observed between the CV values of different fertility indices and CV of IFI: CV values of pH, alkali-hydrolyzable nitrogen, and available potassium were positively correlated with CV of IFI, with the correlation for available potassium reaching a significant level (P<0.05); while CV values of organic matter and available phosphorus were negatively correlated with CV of IFI. although these correlations were not significant. A fitting equation was established to predict the variation in soil fertility based on the variation of soil fertility indicators: CV_IFI =0.606×CV_pH+0.06×CV_OM-0.109×CV_AN-0.079×CV_AP+0.178×CV_AK+9.884, with an R² of 0.478. Furthermore, the fitted CV of the IFI was linearly correlated with the actual CV of the IFI (y = 0.48x + 11.49), with an R² of 0.48, a root mean square error (RMSE) of 2.73, and a normalized root mean square error (n-RMSE) of 12.37%, indicating high stability. This relationship can be used to approximately predict the variability of soil fertility based on the variation of soil fertility indicators.