Objective The Cox maze III procedure has excellent long-term efficacy in curing atrial fibrillation. It has not been widely practiced because it is technically challenging and requires prolonged cardiopulmonary bypass. The aim of this study was to examine a simplified Cox maze III procedure that uses bipolar radiofrequency energy as an ablative source. Methods Beginning January 2002, a total of 40 consecutive patients underwent a modified Cox maze III procedure with bipolar radiofrequency energy. Nineteen had a lone maze procedure and 21 had a maze procedure plus a concomitant operation. One month after the operation, the first 8 patients were investigated with high-resolution magnetic resonance imaging. Patients were followed up monthly with clinical examination and electrocardiography. Results There was no operative deaths. The crossclamp times were 47 ± 26 minutes for the modified lone Cox maze III procedure and 92 ± 37 minutes for the Cox maze III procedure plus concomitant procedures. These were significantly shorter than our previous times for the traditional Cox maze III procedure (93 ± 34 minutes and 122 ± 37 minutes, respectively, P < .05). Follow-up magnetic resonance imaging showed no evidence of pulmonary vein stenosis, and atrial contractility was preserved in all patients. There were no late strokes. At 6-month follow-up, 91% of patients (21/23) were in sinus rhythm. Conclusions Bipolar radiofrequency ablation can be used to replace the surgical incisions of the Cox maze procedure. This energy source did not result in pulmonary vein stenosis. The modification of the Cox maze III procedure to use bipolar radiofrequency ablation simplified and shortened this procedure without sacrificing short-term efficacy.