Abstarct: Quantitative Structure-Retention Relationship (QSRR) is one of the most effective and widely applied approaches for predicting chromatographic properties of compounds baxsed on their molecular structure. In this study, QSPR/QSRR models were developed to predict the linear temperature-programmed retention index (LTPRI) of a diverse set of organic compounds, including ketones, esters, trimethylsilyl (TMS) derivatives of phenols, aliphatic acids, and aromatic acids. The dataset comprised 120 compounds with experimental RI values obtained from a reliable source. Molecular descxriptors were calculated using Dragon software and subsequently reduced through variable selection techniques, including Stepwise Regression (SW), Genetic Algorithm (GA), Firefly Algorithm (FF), and Random Forest (RF), to identify optimal and low-collinearity subsets. Nonlinear relationships between the selected descxriptors and LTPRI were modeled using Artificial Neural Networks (ANN). Model performance was evaluated using an external test set along with various validation procedures. The results indicated that the GA-ANN model exhibited superior predictive performance compared to the other approaches. Nevertheless, the SW-ANN model achieved comparable accuracy using a significantly smaller number of descxriptors, resulting in a simpler, more interpretable, and computationally efficient model. Y-randomization tests and applicability domain analysis confirmed the absence of chance correlations and the genuine validity of the developed models. The findings of this study validate the effectiveness of ANN-baxsed QSPR/QSRR models for LTPRI prediction and demonstrate that stepwise regression can serve as an efficient strategy for descxriptor selection without substantial loss in predictive accuracy. The proposed approach provides a rapid, reliable, and cost-effective tool for estimating LTPRI in dual-phase chromatographic systems (n-hexane/nitromethane) and can significantly reduce the need for extensive and time-consuming experimental measurements.