Background: To evaluate the impact of radiation dose escalation on overall survival (OS) in patients with non-metastatic esophageal squamous cell carcinoma (ESCC) treated with radical radiotherapy. Methods: The clinical data of ESCC patients treated with three-dimensional (3D) radiotherapy alone or chemoradiotherapy were collected from multiple institutes and retrospectively analyzed. Patients who received radiation dose >= 40 Gy were included. Radiation dose as a continuous variable was entered into the Cox regression model by using penalized spline regression to allow for a nonlinear relationship between radiation dose and OS to be identified. Patients were stratified into five groups according to EQD2. The Kaplan-Meier method was used to assess the OS in different dose groups. Univariate and multivariate analyses were performed to evaluate the factors associated with OS. Results: A total of 2,469 patients were included from 10 institutes across China. The median follow-up time was 58.3 months [95% confidence interval (CI): 56.4-60.2 months]. The median OS and PFS time were 24.3 months (95% CI: 22.5-26.2 months) and 18.0 months (95% CI: 16.4-19.6 months), respectively. The risk of death decreased sharply with a dose up to 60 to 62 Gy, before increasing slightly after the dose was elevated beyond 62 Gy. Multivariate analysis indicated that the chance of death was significantly decreased in patients who received radiotherapy doses of 60-62 Gy [P=0.028, hazard ratio (HR) 0.85, 95% CI: 0.73-0.98)], compared with those who received radiotherapy doses of 40-60 Gy. Conclusions: Our results reveal radiation dose is a significant prognostic factor of survival for ESCC patients. Higher radiation dose contributes to much more favorable survival outcomes for ESCC patients receiving radical radiotherapy by modern techniques, and 60 Gy or above might be the most optimal radiation dose.