Voltage-guided substrate ablation following pulmonary vein isolation (PVI) improves atrial fibrillation (AF) ablation outcomes. However, by setting an upper voltage cutoff of 0.5 mV during sinus rhythm (SR) to guided substrate ablation using electroanatomic voltage mapping (EAVM), mildly affected low-voltage area (maLVA) may be undetected. We sought to determine the optimal bipolar voltage cutoff to identify maLVA, its electrogram complexity, and the implication on ablation outcome. Left atrial (LA) EAVMs were obtained in patients without AF and structural heart disease (control) to devise a voltage cutoff to identify maLVA. Subsequently, we investigated 100 patients without low-voltage area (LVA) of < 0.5 mV who underwent PVI alone. In our 6 control cohorts, 95% of LA regional bipolar voltage was > 1.17 mV. maLVA, defined as <1.1 mV, was present in 43% of AF patients, associated with higher prevalence of abnormal electrograms (44.1% vs. 4.4%, P < 0.001). During a median of 2.4 years, patients with maLVA had higher recurrence rate (Log-rank P < 0.001), and maLVA was an independent predictor for recurrence in a multivariate analysis (hazard ratio [HR] 3.944; 95% confidence interval [CI] 1.292-12.042; P = 0.016). A control-derived LA voltage cutoff of <1.1 mV for EAVM in SR reveals maLVA, harboring abnormal electrograms, as an independent predictor for recurrences after PVI alone in patients without LVA (< 0.5 mV). Adjunctive maLVA-guided substrate ablation targeting mildly remodeled and potentially arrhythmogenic LA substrate may further improve the long-term outcome of AF ablation.