BACKGROUND: Robot-assisted minimally invasive esophagectomy (RAMIE) is an effective but technically demanding procedure. The learning curve of RAMIE has been studied to help guide training and to ensure its safe implementation. METHODS: We retrospectively analyzed the first 83 consecutive patients with thoracic esophageal cancer who underwent robot-assisted minimally invasive Mckeown esophagectomy (RAMIE-MK) between May 2021 and August 2023, all performed by a single surgeon. A cumulative sum (CUSUM) analysis was applied to generate the learning curve of RAMIE-MK, based on total operation time. RESULTS: The learning curve was divided into two phases based on the CUSUM analysis: Phase I, the initial learning phase (cases 1-27) and Phase II, the proficiency phase (cases 28-83). When comparing the proficiency phase with the initial phase, we observed a significant decreased trends in total operation time (329.6 ± 71.0 min vs 221.3 ± 33.5 min, P＜0.001). No significant differences were found in other clinicopathological characteristics. CONCLUSION: For a surgeon experienced in open and thoracolaparoscopic esophagectomy, and who also received systematic robot-assisted thoracic surgery training on animals, a total of 27 cases were required to gain technical proficiency in RAMIE-MK.