Background and Aims: Endoscopic submucosal dissection (ESD) seems to be a reasonable option for gastrointestinal subepithelial lesions (SELs) localized within the submucosa. Indications for ESD include small neuroendocrine tumors (NETs) and indeterminate SELs. However, the prospective data regarding ESD and surveillance remain unclear. This study was performed to prospectively investigate the outcomes of ESD, including organ-specific outcomes and the mid-term prognosis. Methods: This prospective multicenter study included 57 patients who underwent ESD for SELs localized within the submucosa [definite NETs (n = 42) and indeterminate SELs (n = 15)]. The efficacy and safety of ESD were evaluated in the whole cohort and in subgroups (NETs and indeterminate SELs). All patients were followed up. Results: The rates of en bloc resection, curative resection, and complications were 98.2%, 66.7%, and 7.7% for the overall population (n=57); 100%, 61.9%, and 2.4% for NETs (n=42); and 93.3%, 80.0%, and 20.0% for indeterminate SELs (n=15), respectively. The rates of curative resection for NETs were poorer in the stomach (20%, n=5) and duodenum (33%, n=3) than in the rectum (71%, n=34). Including 11 of 16 patients with NETs who underwent a conservative approach resulting in non-curative resection, no patients developed tumor recurrence during the follow-up period (median, 24.5 months; range, 1–60 months). ESD followed by surveillance demonstrated acceptable mid-term outcomes for non-curative NETs. Conclusions: ESD can be an efficient therapy for SELs localized within the submucosa. However, gastric and duodenal ESD for NETs may be limited in terms of its curative and technical aspects. Clinicians should be aware of the potential complications of ESD for indeterminate SELs.