Duodenal endoscopic submucosal dissection (ESD) for superficial non-ampullary duodenal epithelial tumors has a significant incidence rate of delayed perforation. Although several methods have been proposed to prevent delayed perforation, the most appropriate methods remain unclear. Currently, there is no appropriate animal model to validate methods for preventing duodenal delayed perforation. This study aimed to establish an in-vivo porcine delayed perforation model after duodenal submucosal dissection. Two porcine models underwent either ESD or surgical submucosal dissection. In the surgical dissection model, an inverted duodenal mucosa was resected with electrosurgical energy. In the ESD model, a gauze was placed behind the duodenum with grasped transverse part to improve endoscopic maneuverability. The mucosal defects after dissection were treated with omental coverage without suture in both models. All models were euthanized 0-5 days after procedure. Body weight; resection size; procedure dissection time; presence of intraoperative perforation and delayed perforation; and adhesion score were assessed. There were no significant differences in body weight and adhesion score between the two models. Resection size was significantly larger in the surgical dissection models than in the ESD models (19 mm vs 14.3 mm, P < 0.01). Procedure time was significantly longer in the ESD models than in the surgical models (45.2 minutes vs 4.5 minutes, P < 0.01). Delayed perforation rates in the surgical dissection models and the ESD models were 0% (0/5) and 100% (5/5), respectively (P < 0.01). This study indicated that our in-vivo porcine duodenal ESD model is beneficial to evaluate a prevention strategy for delayed perforation. © 2020 Japan Gastroenterological Endoscopy Society.