Summary Background Bariatric surgery leads to a higher remission rate for type 2 diabetes mellitus than non‐surgical treatment. However, it remains unsolved which surgical procedure is the most efficacious. This network meta‐analysis aimed to rank surgical procedures in terms of diabetes remission. Methods and findings We electronically searched for randomized controlled trials in which at least one surgical treatment was included among multiple arms and the diabetes remission rate was included in study outcomes. A random‐effects network meta‐analysis was performed within a frequentist framework. The hierarchy of treatments was expressed as the surface under the cumulative ranking curve value. Results of the analysis of 25 eligible randomized controlled trials that covered non‐surgical treatments and eight surgical procedures (biliopancreatic diversion BPD, BPD with duodenal switch, Roux‐en Y gastric bypass, mini gastric bypass mini‐GBP, laparoscopic adjustable gastric banding, laparoscopic sleeve gastrectomy, greater curvature plication and duodenal‐jejunal bypass) showed that BPD and mini‐GBP had the highest surface under the cumulative ranking curve values among the eight surgical treatments. Conclusion Current network meta‐analysis indicated that BPD or mini‐GBP achieved higher diabetes remission rates than the other procedures. However, the result needs to be interpreted with caution considering that these procedures were in the minority of bariatric surgeries.