Transcranial magnetic stimulation (TMS) provides a method to noninvasive excitation of the facial nerve in its intracranial segment close to the internal acoustic meatus. Thus, the site of facial nerve activation with TMS is proximal to or within the site of the lesion in Bell's palsy. To evaluate the prognostic capability of TMS in unilateral Bell's palsy we examined 137 patients with this method, and compared the results with electroneuronography (ENoG). Within 0-4 days from the onset of palsy, the patients with elicitable TMS responses recovered better than those in whom TMS responses were not elicitable. If TMS was performed 5-9 days or 10-28 days after the onset of palsy, it did not provide any prognostic information. Based on amplitude side-to-side differences, ENoG did not contribute prognostic information during the first 9 days from the onset of palsy. Later on, 10-28 days after the onset of palsy, ENoG showed an increased capability to discriminate the patients with poor prognosis. Thus, elicitable facial motor response with TMS predicts good prognosis of Bell's palsy at an early stage whereas poor response with ENoG predicts less favorable prognosis at a later stage.