Background: Remote-access robotic thyroid surgery enables avoiding a visible scar on the neck and allows precise manipulation through a magnified surgical view. The retroauricular approach has many advantages.This study aimed to evaluate the learning curve for robotic retroauricular thyroidectomy using cumulative sum analysis. Methods:The medical records of 36 patients who underwent robotic retroauricular thyroidectomy between 2018 and 2021 were retrospectively reviewed. The clinical features and surgical outcomes were analyzed; the learning curve was evaluated using the cumulative sum analysis.Results: The learning curve using cumulative sum analysis was divided into two phases based on 15 cases: phase I (first 15 cases) and phase II (remaining 21 cases). The total operation time was significantly shorter in phase II than that in phase I (161.9±23.4 vs. 199±41.0 min, P=0.002). The flap dissection and docking time (77.1±14.3 vs. 90.0±21.5 min, P=0.037) and console time (36.5±16.2 vs. 50.3±17.8 min, P=0.020) were significantly shorter in phase II than that in phase I. There was no significant difference between the two phases in the total amount of drainage, duration of hospital stay, and complications after the surgery. Conclusions:The learning curve for robotic retroauricular thyroidectomy demonstrates that the operation time decreased rapidly after 15 cases. Proficiency in docking and manipulating the instruments accelerate the learning curve.