Current guidelines for the optimal ablation strategy for persistent atrial fibrillation(PerAF) remain unclear. While our previous RCT confirmed the favorable prognosis of aggressive ablation, real-world evidence is still lacking. Among 4,833 PerAF patients undergoing catheter ablation at 10 centers, two groups were defined: Regular Ablation(PVI-only or PVI plus anatomical ablation) and Aggressive Ablation(anatomical plus electrogram-guided ablation), with 1,560 patients each after propensity score(PS) matching. The primary endpoint was 12-month AF/atrial tachycardia(AT) recurrence-free survival off anti-arrhythmic drugs after a single procedure. Additional PS matching was performed within the regular group between PVI-only and anatomical ablation(n=455 each). Furthermore, anatomical ablation from the regular group was independently matched with aggressive ablation (n=1,362 each). At 12 months, the Aggressive Group showed superior AF/AT-free survival(66.2% vs. 59.3%, p<0.001; HR 0.745), similar AT recurrence(12.0% vs. 11.3%, p=0.539), and significantly higher procedural AF termination(67.0% vs. 21.0%, p<0.001) than Regular Group. Moreover, patients with AF termination had improved AF/AT-free survival(72.3% vs. 55.2%, p<0.001). Safety endpoints did not differ significantly between the two groups. Both the ablation outcomes and AF termination rate showed increasing trends with the extent of ablation aggressiveness but declined with extremely aggressive ablation. After additional PS matching, within the regular group, no statistical differences were observed though AF/AT-free survival in the anatomical group was slightly higher than the PVI-only group (60.7% vs. 55.6%, p=0.122); while aggressive ablation showed improved AF/AT-free survival compared to anatomical ablation alone from regular group (67.5% vs. 59.9%, p<0.001). Aggressive ablation achieved more favorable outcomes than regular ablation, and moderately aggressive ablation may be associated with better clinical outcomes. AF termination is a reliable ablation endpoint.