Purpose: Laparoscopic gastrectomy is a widely accepted surgical technique. Recently, robotic gastrectomy has been developed, as an alternative minimally invasive surgical technique. This study aimed to evaluate the question of whether robotic gastrectomy is feasible and safe for the treatment of gastric cancer, due to its learning curve. Materials and Methods: We retrospectively reviewed the prospectively collected data of 100 consecutive robotic gastrectomy patients, from November 2008 to March 2011, and compared them to 282 conventional laparoscopy patients during the same period. The robotic gastrectomy patients were divided into 20 initial cases; and all subsequent cases; and we compared the clinicopathological features, operating times, and surgical outcomes between the three groups. Results: The initial 20 robotic gastrectomy cases were defined as the initial group, due to the learning curve. The initial group had a longer average operating time ($242.25{pm}74.54$ minutes vs. $192.56{pm}39.56$ minutes, P>0.001), and hospital stay ($14.40{pm}24.93$ days vs. $8.66{pm}5.39$ days, P=0.001) than the experienced group. The length of hospital stay was no different between the experienced group, and the laproscopic gastrectomy group ($8.66{pm}5.39$ days vs. $8.11{pm}4.10$ days, P=0.001). The average blood loss was significantly less for the robotic gastrectomy groups, than for the laparoscopic gastrectomy group ($93.25{pm}84.59$ ml vs. $173.45{pm}145.19$ ml, P<0.001), but the complication rates were no different. Conclusions: Our study shows that robotic gastrectomy is a safe and feasible procedure, especially after the 20 initial cases, and provides a satisfactory postoperative outcome.