Objective: To explore the efficacy and safety on upper gastrointestinal submucosal tumors treated by endoscopic submucosal excavation (ESE), endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD). Methods: A retrospective analysis was conducted on 307 patients with upper gastrointestinal submucosal tumors. Patients treated with ESE, EMR, and ESD were included in ESE group (n=98), EMR group (n=105) and ESD group (n=104) respectively. Perioperative indicators, efficacy, gastric function and postoperative complications were compared among the three groups. Results: The surgical time in EMR group was shorter than that in ESE and ESD groups (P<0.05), and the surgical time was shorter in ESE group than that in ESD group (P<0.05), and the hospitalization time was shorter in ESE and EMR groups (P<0.05). The tumor complete resection rate, first exhaust time, first defecation time, pepsinogen I, pepsinogen II, and gastrin-17 revealed no statistical differences among the three groups (P>0.05). The recurrence rate in ESE group was lower compared to EMR group (P<0.05). The incidence rates of postoperative complications in EMR group were lower than those in ESE group and ESD group (P<0.05). Conclusion: ESE, EMR and ESD are effective in the treatment of upper gastrointestinal submucosal tumors. EMR has faster postoperative recovery, fewer complications but higher recurrence rate, and it is recommended for submucosal tumors. ESE achieves complete resection and safety, and is suitable for tumors involving the deep or shallow muscular layer of the submucosa. ESD can reduce the risks of residual and recurrence of deep muscle layer or muscularis propria tumors.