Abstract Background This study aims to assess the learning curve of robotic distal gastrectomy (RDG) and robotic total gastrectomy (RTG) for gastric cancer. Methods Data on consecutive patients who underwent robotic gastrectomy for gastric cancer by five surgeons between March 2010 and August 2019 at two high-volume institutions were collected. The learning curve was determined based on the analyses of operation time and postoperative complications within 30 days. Cumulative sum analysis (CUSUM) and risk-adjusted-CUSUM (RA-CUSUM) were applied to identify the turning points (TPs). Results A total of 899 consecutive patients were included. The mean number of patients needed to overcome the learning curve for operation time of RDG and RTG were 22 and 20, respectively. The number of patients needed to overcome the learning curve for postoperative complications after RDG and RTG were 23 and 18, respectively. The surgical outcomes in the post-TP group were better than in the pre-TP group and improved as surgeons’ experience increased. Also, increased case numbers in RDG promoted the RTG learning process. Conclusion The present study demonstrated a substantial influence of surgical cumulative volume on improved surgical outcomes in robotic gastrectomy. Increased experience in RDG may help surgeons to achieve proficiency faster in RTG.