The enantiomeric resolution of chiral compounds using HSA by means of affinity EKC (AEKC)–partial filling technique is the result of a delicate balance between different experimental variables such as protein concentration, running pH (background electrophoretic buffer, protein and compound solutions) and protein solution plug length. In this paper multivariate optimization approaches for chiral separation of four basic drugs (alprenolol, oxprenolol, promethazine and propranolol) using HSA as chiral selector in AEKC–partial filling technique are studied. The experimental conditions to achieve maximum resolution are optimized using the Box–Behnken experimental design. Partial least squares and pareto charts are used to analyse the main effects on the resolution. The experimental resolutions observed for all compounds studied in optimum conditions agree with the estimated values based on response surface models. The results obtained show that the range of experimental conditions that provided enantioresolution narrows as hydrophobicity of analytes decreases. This fact can be explained by assuming that hydrophobicity controls the interaction of basic compounds with HSA.