Background Endotracheal tube size can be predicted according to ultrasound measurement of subglottic airway diameter. The learning curve for this method is not yet established. The aim was to evaluate the learning curve of anesthesiology residents in ultrasound measurement of subglottic airway diameter for prediction of endotracheal tube size using cumulative sum analysis. Methods Sixteen anesthesiology residents measured transverse subglottic airway diameter in children undergoing general anesthesia with cuffed endotracheal intubation. Each resident performed 30 ultrasound examinations. Primary outcome was the successful prediction of endotracheal tube size according to ultrasound measurement. Cumulative sum analysis was performed with acceptable and unacceptable failure rates set as 20 and 40%, respectively. Results Ten out of 16 residents (62.5%) were deemed successful as they were able to pass lower decision boundary, whereas six residents' CUSUM scores were between the decisions lines deeming them indeterminate. The overall success rate for determining the correct endotracheal tube size was 77.5%. Median number of attempts to cross lower decision boundary was 29 with minimum of 18 and maximum of 29 attempts among successful residents. Conclusion Learning curves constructed with cumulative sum analysis in this study showed that only 62.5% of residents were able to correctly predict cuffed endotracheal tube size with 80% success rate. Considerable variability in achieving competency necessitates objective follow-up of individual improvement.