The adjunctive approach is still unknown for atrial fibrillation (AF), which cannot be terminated after pulmonary vein isolation (PVI). We hypothesized that the driver ablation plus PVI was superior to PVI alone. A total of 98 patients with paroxysmal AF were enrolled in this study and were divided into two groups, with one group undergoing PVI (n = 49) and the other group undergoing PVI + driver ablation (n = 49). The driver regions were defined as clusters of bipolar electrograms that displayed spatial dispersion spread over mean AF cycle length at a minimum of three adjacent bipolars of a PentaRay catheter. During the procedure, the most prominent driver regions before PVI were the roof (n = 27; 55.1%), PV antrum (n = 23; 46.9%), and the inferoposterior wall (n = 11; 22.4%). PVI can eliminate all drivers at PV antrum, but only terminate 30.4% of AF in the driver group. The AF termination rate in the driver ablation group was significantly higher than that in conventional ablation (93.9% vs 40.6%; P < 0.001). The rate of freedom from atrial tachyarrhythmia episodes by a single procedure at 6 months was significantly higher in the driver group than in the conventional group (91.6% vs 72.4%; P = 0.02). The present method is effective for AF driver identification. It guided ablation adjunctive to PVI increasing the rate of AF termination and improving the outcomes in patients with paroxysmal AF.