Objective To explore the establishment of a risk nomogram prediction model for radiation esophagitis in patients with esophageal cancer. Method A total of 94 patients with esophageal cancer who underwent radiation therapy in Department of VIP Medical Services of the Cancer Hospital Chinese Academy of Medical Sciences and Peking Union Medical College from October 2017 to July 2022 were selected as the study objects. They were divided into two groups according to whether radiation esophagitis occurred during follow-up including 53 patients in the radiation esophagitis group and 41 patients in the non -radiation esophagitis group. Multivariate Logistic regression analysis was used to screen the independent risk factors of radiation esophagitis in patients with esophageal cancer and the nomogram model was established on this basis. Result There were statistically significant differences in surgical conditions nutritional support tumor location TNM stage and hypoproteinemia between radiation esophagitis group and non-radiation esophagitis group P< 0. 05 . Multivariate Logistic regression analysis showed no surgery OR = 3. 152 95% CI = 1. 070-4. 471 no nutritional support OR = 2. 330 95%CI = 1. 268-3. 544 tumor location in the neck OR = 3. 380 95% CI = 1. 197- 10. 551 and TNM stage ≥ Ⅲ OR = 2. 625 95% CI = 1. 387 - 5. 554 were independent risk factors for radiation esophagitis. Based on the results of multivariate Logistic regression analysis a nomogram prediction model was established and receiver operating characteristic curve ROC was used to evaluate the prediction efficiency of radiation esophagitis. The results showed that the area under the curve AUC was 0. 730. Conclusion Radiation esophagitis may occur in some patients with esophageal cancer after radiotherapy. The independent risk factors for radiation esophagitis include no operation no nutritional support neck location of the tumor and TNM stage ≥Ⅲ. The establishment of a quantitative nomogram prediction model based on the above risk factors has good efficacy and clinical application value.