It remains unclear whether AF is maintained by rotor. We evaluated the significance of rotor during atrial fibrillation (AF). Prevalence, location, and stability of rotational reentry (RR) in the left atrium were clarified by endocardial non-contact mapping in 66 AF patients. RR was classified into three categories: RR continued at stable site (Stable-RR), RR observed intermittently at the same site (Intermittent-RR), and RR observed at different locations (Different-RR). Catheter ablation was performed in a stepwise fashion (linear roof lesion and complex fractionated atrial electrogram ablation following pulmonary vein isolation) until AF termination and elucidated the consequence of radiofrequency lesion delivered within RR site on AF termination and recurrence. One hundred and nineteen RRs were observed. There were 54 patients with RR (RR Group) and 22 patients without RR (Non-RR Group). Prevalence of Different-RR (n = 81) was significantly higher than Stable-RR (n = 16, p < 0.001) and Intermittent-RR (n = 22, p < 0.001). The intervals involved in RR occupied only 22.4% of total activation time. There was no significant difference in the prevalence of AF termination nor AF/atrial tachycardia recurrence between RR and non-RR Groups (46 vs. 9 patients, p = 0.317, and 13 vs. 1 patients, p = 0.271) and between patients in whom radiofrequency lesion was involved in RR and those was not (24 vs. 22 patients, p = 0.210, and 6 vs. 7 patients, p = 0.506). In conclusion, most RRs were observed transiently and often shifted its locations. Radiofrequency lesion delivered within RR site did not correlate with AF termination nor recurrence, suggesting that RR is not a driving source during AF.