The retention behaviour of 22 selected imidazoline drugs and derivatives was investigated on α₁-acid glycoprotein (AGP) column using Sørensen phosphate buffer (pH 7.0) and 2-propanol as organic modifier. Quantitative Structure-Retention Relationships (QSRR) models were built using extrapolated logk_w values as well as isocratic retention factors (logk₅, logk₈, logk₁₀, logk₁₂, logk₁₅ obtained for 5%, 8%, 10%, 12%, and 15%, of 2-propanol in mobile phase, respectively) as dependant variables and calculated physicochemical parameters as independant variables. The established QSRR models were built by stepwise multiple linear regression (MLR) and partial least squares regression (PLS). The performance of the stepwise and PLS models was tested by cross-validation and the external test set prediction. The validated QSRR models were compared and the optimal PLS-QSRR model for logk_w and each isocratic retention factors (PLS-QSRR(logk₅), PLS-QSRR(logk₈), PLS-QSRR(logk₁₀), MLR-QSRR(logk₁₂), MLR-QSRR(logk₁₅)) were selected. The QSRR results were further confirmed by Linear Solvation Energy Relationships (LSER). LSER analysis indicated on hydrogen bond basicity, McGowan volume and excess molar refraction as the most significant parameters for all AGP chromatographic retention factors and logk_w values of 22 selected imidazoline drugs and derivatives.