In this paper, Kováts retention indices determined on stationary phases with chemically bonded copper complexes were correlated with molecular structural parameters for aromatic compounds. Principal component regression (PCR) was applied to extract principal components from the set of 13 descriptors compiled in 5 theoretical models. Extracted components were used to model theoretical retention indices. Significant correlations were found among the retention indices of these compounds and, among others: molecular surface and molecule area, boiling point, HOMO and LUMO energies, dipole moment, dielectric energy, and double bond count. These correlations provide insights into the mechanism of chromatographic retention at the molecular level for the packings and the compounds under study.