Abs are very efficient drugs， ∼70 of them are already approved for medical use， over 500 are in clinical development， and many more are in preclinical development. One important step in the characterization and protection of a therapeutic Ab is the determination of its cognate epitope. The gold standard is the three-dimensional structure of the Ab/Ag complex by crystallography or nuclear magnetic resonance spectroscopy. However， it remains a tedious task， and its outcome is uncertain. We have developed MAbTope， a docking-based prediction method of the epitope associated with straightforward experimental validation procedures. We show that MAbTope predicts the correct epitope for each of 129 tested examples of Ab/Ag complexes of known structure. We further validated this method through the successful determination， and experimental validation (using human embryonic kidney cells 293)， of the epitopes recognized by two therapeutic Abs targeting TNF-α: certolizumab and golimumab.