Direct pharmacokinetic/pharmacodynamic relationships for otamixaban were investigated after rising doses in healthy subjects using mixed‐effect modeling. Activated partial thromboplastin time, prothrombin time, dilute prothrombin time, and Russell's viper venom—induced clotting time (RVVT) related linearly, whereas Heptest clotting time (HCT) followed a sigmoidal E max model. The pharmacokinetic/pharmacodynamic response (slope) and their corresponding interindividual variability (seconds per ng/mL, [% coefficient of variation]) were 0.263 (29%) for Russell's viper venom—induced clotting time, 0.117 (10%) for dilute prothrombin time, 0.058 (19%) for activated partial thromboplastin time, and 0.021 (11%) for prothrombin time. For Heptest clotting time, the parameter estimates with their corresponding interindividual variability (% coefficient of variation) were 71 ng/mL (30%) for EC 50 , 186 seconds (64%) for E max , and 17 seconds (16%) for E 0 . The model predicted otamixaban plasma concentrations to double the clotting times that were close to those observed. These pharmacokinetic/pharmacodynamic relationships, together with the predictable pharmacokinetics, allowed the anticoagulant effect at given doses of otamixaban to be foreseen in healthy subjects.