This study is to assess pharmacokinetic (PK) sampling time schedules and trial size requirements of drug-drug interaction (DDI) studies for CYP2C9, based on S-warfarin population PK models. S-warfarin plasma concentrations from eight DDI studies were utilized to develop S-warfarin population PK models. Optimal PK sampling times were obtained that minimized mean squared error of geometric mean of the area under the concentration-time curve (AUC0-∞). The powers and type I error rates of testing the equivalences of the geometric means of AUC0-∞ only and AUC0-∞ and maximum concentration (Cmax), jointly, were assessed via simulation for two-by-two crossover designs. The results were compared to those from three bioequivalence sample size calculation methods. Two-compartment population PK models with first order absorption were established for non-Asian and Asian subjects. The optimal PK sampling times of size 17 per individual per period were found to be 0.0, 0.5, 1.0, 2.0, 4.0, 6.0, 10.0, 12.0, 16.0, 24.0, 36.0, 48.0, 60.0, 72.0, 96.0, 120.0, 144.0 h post a single oral dose of 25 mg warfarin. For non-Asian subjects, the minimum numbers of subjects required per trial with the optimal PK sampling schedule to achieve 80% power and 5% type I error rate, ranged from 6 to 19 for the equivalence of AUC0-∞ and Cmax jointly. It has been demonstrated that appropriately selected PK sampling time points can greatly increase the corresponding power of the study without increasing the number of subjects, especially when the true ratio is near the default bioequivalence boundary (0.8-1.25). © Springer Science+Business Media, LLC 2012.