Cessation of oral anticoagulation (OAC) is common after the first 3 months of catheter ablation of atrial fibrillation (AF); however, thromboembolic risk has not been defined in patients with and without AF recurrence (RAF vs. NRAF) post ablation. We identified 796 patients who discontinued OAC at 3 months post AF ablation from January 2015 to May 2018 in our center. Regular follow-up was performed to detect RAF, collect medication management and thromboembolic and major bleeding events. CHA2DS2-VASc score was 1.79±1.50; 547 (68.7%) patients were at intermediate and high risk (i.e. CHA2DS2-VASc score ≥1 in male patients, or ≥2 in female patients); 169 (21.2%) were RAF. During 29.2±12.2 months follow-up, the incidence rate of thromboembolism was 1.62 per 100 patient-year (7 in 431 years) in RAF, 0.33 per 100 patient-year (5 in 1503 years) in NRAF. After adjusting for potential confounding factors, RAF was associated with more 3.5-fold higher rate of thromboembolism compared with NRAF (adjusting HR, 4.488; 95%CI, 1.381-14.586). Rate of thromboembolism was even higher in patients with intermediate and high risk (2.16 per 100 patient-year [7 in 323 years] versus 0.38 per 100 patient-year [4 in 1043 years], aHR, 5.807; 95%CI, 1.631-20.671). In multivariate logistic regression analysis, RAF was the only independent predictor of thromboembolism (4.837 [1.498-15.621], P=0.008). In conclusion, cessation of OAC in NRAF may be reasonable, especially for patients with the contraindications for continuing OAC; however, cessation of OAC appeared unsafe in RAF with a high-risk stroke profile because of high incidence rate of thromboembolism.