Background Ethambutol (EMB) is a first-line antituberculosis drug of which renal elimination is the primary route of elimination. EMB has been proven to be a substrate of OCT2. The EMB concentration significantly increases in renal failure patients. It is of importance to evaluate renal function, OCT2 genotypes as well as the interplay among them on EMB pharmacokinetics. Therefore, we aimed to develop a physiologically based pharmacokinetic (PBPK) model implying those variables to understand more in detail and predict EMB exposure in individual patients. Methods The effect of OCT2 genetic variants on OCT2-mediated uptake of EMB was estimated using in vitro stably transfected HEK293 cells with OCT2 wild type and variants. A PBPK model for EMB was constructed in Simcyp (V.17) and validated against clinical data. The validated model was used to predict the effect of OCT2 genotypes and renal impairment on EMB disposition. The predicted EMB exposure in “virtual twins” tuberculosis patients was compared with clinical data. Results OCT2-genetic variants showed a significant decrease in OCT2-mediated uptake of EMB in vitro compared to wild type. EMB plasma concentration predicted by EMB PBPK model compared well with published clinical data. OCT2-T199I increases EMB AUCinf by 42% and decreases renal clearance (CLr) by 46% compared to that of wild type. An increased AUCinf ratio of 1.3 and 1.5 among GFR 30-60 and GFR 15-30 groups is predicted, with significantly reduced CLr than healthy. Most importantly, terminally impaired GFR and OCT2-T199I combined caused approximately 2 fold increase of EMB AUCinf, reduced about 4.5 fold CLr than healthy subjects. The model predicted quite good observed data in "virtual twin" tuberculosis patients. Conclusion To the best of our knowledge, this is the first PBPK model showing a significant effect of renal impairment and transporter genotypes on EMB pharmacokinetics, a roadmap towards personalized dose estimation.