Catheter ablation (CA) is a well-established therapy for rhythm control in atrial fibrillation (AF). However, CA outcomes for persistent AF remain unsatisfactory because of the high recurrence rate despite time-consuming efforts and the latest ablation technology. Therefore, the selection of good responders to CA is necessary. Surface electrocardiography (sECG)-based complexity parameters were tested for the predictive ability of procedural termination failure during CA and late recurrence of atrial arrhythmias (AA) after CA. A total of 130 patients with nonparoxysmal AF who underwent CA for the first time were investigated. A 10-second sECG of 4 leads (leads I, II, V 1 , and V 6 ) was analyzed to compute the fibrillatory wave amplitude (FWA), dominant frequency (DF), spectral entropy (SE), organization index (OI), and sample entropy (SampEn). The study endpoints were procedural termination failure during CA and late (≥1 year) AA recurrence after CA. In the multivariate analysis, FWA in lead V 1 and DF in lead I were independent predictors of successful AF termination during CA (P <.05). The optimal cut-off values for FWA in lead V 1 and DF in lead I were 60.38 μV (area under the curve [AUC], 0.672; P = .001) and 5.7 Hz (AUC, 0.630; P = .016), respectively. The combination of FWA of lead V 1 and DF of lead I had a more powerful odds ratio for predicting procedural termination failure (OR, 8.542; 95% CI, 2.938-28.834; P < .001). FWA in lead V 1 was the only independent predictor of late recurrence after CA. The cut-off value is 65.73 μV which was 0.634 of the AUC (P = .009).These sECG parameters, FWA in lead V 1 and DF in lead I, predicted AF termination by CA in patients with nonparoxysmal AF. In particular, FWA in lead V 1 was an independent predictor of late recurrence of AA after CA.