Prediction of recurrences after catheter ablation of atrial fibrillation (AF) remains challenging. We sought to investigate the long-term outcomes after AF catheter ablation. A total of 2221 consecutive patients who underwent catheter ablation for symptomatic AF were included in this study (mean age 55 ± 11 years, 20.3% women, and 59.0% paroxysmal AF). Extensive ablation, in addition to circumferential pulmonary vein isolation, was more often accomplished in patients with non-paroxysmal AF than in those with paroxysmal AF (87.4% vs. 25.3%, p < 0.001). During a median follow-up of 54 months, sinus rhythm (SR) was maintained in 67.1% after index procedure. After redo procedures in 418 patients, 83.3% exhibited SR maintenance. Recurrence rates were similar for single and multiple procedures (17.4% vs. 16.7%, p = 0.765). Subanalysis showed that the extent of late gadolinium enhancement (LGE), as assessed by cardiac magnetic resonance, is greater in patients with recurrence than in those without recurrence (36.2 ± 23.9% vs. 21.8 ± 13.7%, p < 0.001). Cox-regression analysis revealed that non-paroxysmal AF (hazard ratio (HR) 2.238, 95% confidence interval (CI) 1.905–2.629, p < 0.001), overweight (HR 1.314, 95% CI 1.107–1.559, p = 0.020), left atrium dimension ≥ 45 mm (HR 1.284, 95% CI 1.085–1.518, p = 0.004), AF duration (HR 1.020 per year, 95% CI 1.006–1.034, p = 0.004), and LGE ≥ 25% (HR 1.726, 95% CI 1.330–2.239, p < 0.001) are significantly associated with AF recurrence after catheter ablation. This study showed that repeated catheter ablation improves the clinical outcomes of patients with non-paroxysmal AF, suggesting that AF substrate based on LGE may underpin the mechanism of recurrence after catheter ablation.