Objective To investigate the role of continuous application of isolated pig esophagus model and synchronous clinical cases in esophageal endoscopic submucosal dissection (ESD) learning curve.Methods The data of isolated pig esophagus model ESD specimen size, dissection time and muscularis propria injury were recorded from January 2017 to December 2019. The data of clinical early esophageal carcinoma continuous ESD cases were collected. One-Way Anova, nonparametric rank sum test and Chi square test were used for statistical analysis.Results In total, we practiced ESD in 36 cases of pig esophagus in vitro. In the first year (2017), the dissection proficiency was significantly lower than that of the second and third year (2018, 2019) (17.2 min/cm² vs 5.6 and 5.8 min/cm², P < 0.05). The muscularis propria injury in the first year (2017) was significantly more than that of the second and third years (2018, 2019) (2.0 vs 0.0 and 0.0, P < 0.05). In total, we performed esophageal ESD in 25 clinical patients. The first year (2017), the dissection proficiency was significantly lower than that of the second and third year (2018, 2019) (39.3 min/cm², vs 16.6 and 12.7 min/cm², P < 0.05), the whole resection rate was 100.0%, the R0 resection rate was 96.0%, and the curative resection rate was 92.0%. Only one case of large ESD had perforation in the early stage, and the total perforation rate was 4.0%, no case of delayed bleeding.Conclusion Endoscopists need to accumulate at least 16 cases of ESD (8 cases of isolated pig esophagus + 8 cases of clinical patients) to master the esophageal ESD technology independently, and for beginners, at least 2 cases of esophageal ESD per month may not be a reasonable load for technical sparsity.