Quantitative structure–property relationship (QSPR) analysis has been carried out to a series of fatty, amino and dicarboxylic acids to model their GC retention times. A genetic partial least square method (GAPLS) was applied as a variable selection tool. Modeling of retention times of these compounds as a function of the theoretically derived descriptors was established by multiple linear regression (MLR) and artificial neural network (ANN). The neural network employed here is a connected back-propagation system with a 3-4-1 architecture. Three topological indices for these compounds, namely, mean information index on atomic composition (AAC), average connectivity index chi-0 (X0A) and total information index of atomic composition (IAC) taken as inputs for the regression models. The results indicate that the GA is a very effective variable selection approach for QSPR analysis. The comparison of the two regression methods used showed that ANN has better prediction ability than MLR. The statistical figure of merits of the two models showed the successful modeling of the retention times with molecular descriptors.