Objective: To demonstrate the learning curve of colposcopic training among gynecologic oncology fellows at King Chulalongkorn Memorial Hospital. Materials and Methods: A retrospective review of colposcopic examinations during June 2014 -April 2017 was performed. The agreement between colposcopic impression (normal vs. low grade lesion vs. high grade lesion) and histopathology (benign vs. HPV/CIN1 vs. CIN2+) was defined as accuracy. Accuracy rate of attending staffs in the institute during the same period was used as the reference level. Cumulative mean proportion of accuracy rate at each consecutive case was plotted in graph and learning curve was generated. Result: Six hundred ninety-one patients were included. Overall accuracy rate of fellows after completion of training was comparable to the attending staffs, which was 68.1% (95% CI 63.4 - 72.4%) versus 68.0% (95% CI 61.9 -73.6%) Normal colposcopic impression correctly predicted normal histopathology in only 30.3% of cases. Colposcopic impression of low-grade and high-grade lesions correctly predicted low-and high-grade intraepithelial lesions on biopsy in 75.9% and 53.4% of cases, respectively. CIN2+ lesions were misdiagnosed in 2.9% of the patients with normal colposcopic impression and 37.2% of the patients with low-grade lesion. The learning curve showed a plateau after 50 cases at around 70% accuracy rate, which was the similar accuracy rate of attending staffs. Conclusion: Colposcopic examination is a procedure that requires comprehensive training. The minimum numbers of 50 colposcopic procedures were required to achieve optimal competency and maintain proficiency.