The need to get a tool for biological parameters estimation of new compounds for clinical applications, supports the postulation of predictive models as an alternative to conventional classical assays being no necessary the use of experimentation in animals. Our main aim in this work is to determine correlations between the logarithm of capacity factors and preclinical pharmacology and therapeutic ef®cacy parameters of barbiturates. The predictive and interpretative capability of quantitative chromatographic retention-biological activity models is supported by the fact that in adequate experimental conditions the solute partitioning into chromatographic system can emulate the solute partitioning into lipid bilayers of biological membranes, which is the basis for drug and metabolite uptake, passive transport across membranes and bioaccumulation. Thirteen barbiturates were included in the study. The RP-HPLC capacity factors of barbiturates were determined using different Brij35 concentrations as micellar mobile phases. Relationships between sixteen biological activities of barbiturates reported in bibliography and retention data are established and their predictive and interpretative ability are evaluated. These logarithmic relationships were sig-ni®cant between preclinical pharmacology parameters and the retention factors of barbiturates; so the regression coef®cients (R 2 ) for ED, EC 50 -LRR and duration of action were 0.97, 0.98 and 0.95, respectively. These relationships were great too for therapeutic ef®cacy parameters; i.e for IC 50 , K i (50% inhibition displaceable [ 14 C]-amobarbital binding to acetylcholine receptor-rich membranes) and K i (for PIP-kinase) (R 2 0.98). The results indicate, the retention of compounds in MLC, which depends on hydrophobic, electronic and steric features of compounds and is measured in ¯ow conditions, is capable to describe and predict in vitro the biological responses of barbiturates. This approach merits further exploration in other classes of compounds.