Introduction Enhanced recovery after surgery (ERAS) protocols have been successfully integrated into peri-operative management of different cancer surgeries such as colorectal cancer. Their value for gastric cancer surgery, however, remains uncertain. Methods A search for randomized and observational studies comparing ERAS versus conventional care in gastric cancer surgery was performed according to PRISMA guidelines. Random-effects meta-analyses with inverse variance weighting were conducted, and quality of included studies was assessed using the Cochrane risk-of-bias tool and Newcastle-Ottawa scale (PROSPERO: CRD42017080888). Results Twenty-three studies involving 2686 patients were included. ERAS was associated with reduced length of hospital stay (WMD-2.47 days, 95% CI − 3.06 to − 1.89, P < 0.00001), time to flatus (WMD-0.70 days, 95% CI − 1.02 to − 0.37, P < 0.0001), and hospitalization costs (WMD-USD$ 4400, 95% CI − USD$ 5580 to − USD$ 3210, P < 0.00001), with consistent results across open and laparoscopic surgery. Postoperative morbidity and 30-day mortality were similar, although a higher rate of readmission was observed in the ERAS group (RR = 1.95, 95% CI 1.03-3.67, P = 0.04). Patients in the ERAS arm had significantly attenuated C-reactive protein levels on days 3/4 and 7, interleukin-6 levels on days 1, and 3/4, and tumor necrosis factor-α levels on days 3/4 postoperatively. Conclusion Compared to conventional care, ERAS reduces hospital stay, costs, surgical stress response and time to return of gut function, without increasing post-operative morbidity in gastric cancer surgery. However, precaution is necessary to reduce the increased risk of hospital readmission when adopting ERAS.