Accurately measuring and predicting biliary excretion would be extremely valuable in evaluating the contribution of biliary excretion to the total systemic clearance, understanding potential mechanisms of hepatobiliary toxicity as well as potentials for drug-drug interactions in drug discovery. In this study, in vivo rat biliary excretion of drug-like molecules was measured using bile duct cannulated rats. Literature biliary excretion data with similar experimental conditions were collected. A predictive quantitative structure-pharmacokinetic relationship (QSPR) model was developed using genetic algorithm guided principal component regression analysis and 2D molecular descriptors. In the derived model, hydrophobicity expressed with calculated distribution coefficients (cLogD) is the most important molecular property correlating biliary excretion. The derived model has been validated using literature data, and should be useful in estimating biliary excretion potentials of molecules in drug discovery.