Abstract: Objective: To classify pulmonary fibrosis severity in connective tissue disease-associated interstitial lung disease (CTD-ILD) patients by integrating visual scoring and quantitative computed tomography (CT). Methods: A total of 96 patients with CTD-ILD and 47 CTD controls without ILD were enrolled. Pulmonary fibrosis was evaluated using the Warrick score and the 3D slicer-based density histogram method. Results: Significant differences were observed in standard deviation (SD), kurtosis, and skewness across the severity groups. Multivariate logistic regression analysis identified HAA as an independent predictor of fibrosis severity. The receiver-operating characterisric curve (ROC) showed that combining all quantitative CT parameters achieved an area under the curve (AUC) of 0.846 (sensitivity=0.739 and specificity=0.885) to distinguish normal/mild cases from moderate/severe cases. Weak positive correlations were observed between Perc10, Perc15, MLA, and RV/TLC. Conclusion: This approach enables a noninvasive and objective assessment of CTD-ILD severity, thereby providing a clinically valuable alternative for patients with limitations in pulmonary function testing, reducing diagnostic and therapeutic costs, and demonstrating a strong potential for widespread implementation in primary healthcare settings.