Quantitative structure–retention relationships (QSRR) offer a powerful approach to predict chromatographic retention times of compounds based on their molecular structures. In this study, we developed a practical and user-friendly QSRR model using multiple linear regression (MLR) to forecast the retention behavior of three classes of illicit veterinary drug additives in food matrices. A total of 95 drugs were analyzed, divided into a training set (62 compounds), a test set (30 compounds), and a real-sample validation set (3 compounds). Molecular descriptors were generated using freely available software tools, including Advanced Chemistry Development (ACD) and the Toxicity Estimation Software Tool (TEST).

The final MLR-QSRR model demonstrated excellent predictive performance, with a strong correlation between observed retention times and selected molecular descriptors (R² = 0.966). Key descriptors influencing retention time included ACDlogP, ALOGP, ALOGP2, Hy, Ui, ib, BEHp1, BEHp2, GATS1m, and GATS2m. Validation through four independent approaches—leave-one-out, k-fold cross-validation, external test set, and real-sample application—confirmed the robustness and reliability of the model.

These findings highlight the potential of QSRR modeling as a valuable analytical tool in food safety, particularly for detecting and monitoring illegal veterinary drug residues. By enabling accurate retention time prediction, this approach supports enhanced screening efficiency in chromatographic workflows and contributes to safeguarding public health.