Radiofrequency ablation (RFA) and endoscopic submucosal dissection (ESD) can potentially be applied for early esophageal squamous cell neoplasia (ESCN); however, no study has directly compared these 2 modalities.We retrospectively enrolled the patients with flat-type “large” (length ≥3 cm extending ≥1/2 of the circumference of esophagus) early ESCNs treated endoscopically. The main outcome measurements were complete response at 12 months, and adverse events.Of a total of 65 patients, 18 were treated with RFA and 47 with ESD. The procedure time of RFA was significantly shorter than that of ESD (126.6 vs 34.8 min; P < 0.001). The complete resection rate of ESD and complete response rate after primary RFA were 89.3% and 77.8%, respectively. Based on the histological evaluation of the post-ESD specimens showed 14 of 47 (29.8%) had histological upstaging compared with the pre-ESD biopsies, and 4 of them had lymphovascular invasion requiring chemoradiation or surgery. After additional therapy for residual lesions, 46 (97.9%) patients in the ESD group and 17 (94.4%) patients in the RFA group achieved a complete response at 12 months. Four patients (8.5%) developed major procedure-related adverse events in the ESD group, but none in the RFA group. In patients with lesions occupying more than 3/4 of the circumference, a significantly higher risk of esophageal stenosis was noted in the ESD group compared with RFA group (83% vs 27%, P = 0.01), which required more sessions of dilatation to resolve the symptoms (median, 13 vs 3, P = 0.04). There were no procedure-related mortality or neoplastic progression in either group; however, 1 patient who received ESD and 1 who received RFA developed local recurrence during a median follow-up period of 32.4 (range, 13–68) and 18.0 (range, 13–41) months, respectively.RFA and ESD are equally effective in the short-term treatment of early flat large ESCNs; however, more adverse events occur with ESD, especially in lesions extending more than 3/4 of the circumference. RFA does not allow for pathology to evaluate the curability after ablation, and thus currently the use for invasive ESCNs should be conservative until longer follow-up studies are available.