Background: Nowadays, there have been increasing studies comparing metformin with insulin. But the use of metformin in pregnant women is still controversial, therefore, we aim to examine the efficiency and safety of metformin by conducting a meta-analysis of randomized controlled trials (RCTs) comparing the effects of metformin with insulin on glycemic control, maternal and neonatal outcomes in gestational diabetes mellitus (GDM). Methods: We used the key words "gestational diabetes" in combination with "metformin" and searched the databases including Pubmed, the Cochrane Library, Web of knowledge, and Clinical Trial Registries. A random-effects model was used to compute the summary risk estimates. Results: Meta-analysis of 5 RCTs involving 1270 participants detected that average weight gains after enrollment were much lower in the metformin group (n = 1006, P = 0.003, SMD = -0.47, 95% CI [-0.77 to -0.16]); average gestational ages at delivery were significantly lower in the metformin group (n = 1270, P = 0.02, SMD = -0.14, 95% CI [-0.25 to -0.03]); incidence of preterm birth was significantly more in metformin group (n = 1110, P = 0.01, OR = 1.74, 95% CI [1.13 to 2.68]); the incidence of pregnancy induced hypertension was significantly less in the metformin group (n = 1110, P = 0.02, OR = 0.52, 95% CI [0.30 to 0.90]). The fasting blood sugar levels of OGTT were significantly lower in the metformin only group than in the supplemental insulin group (n = 478, P = 0.0006, SMD = -0.83, 95% CI [-1.31 to -0.36]). Conclusions: Metformin is comparable with insulin in glycemic control and neonatal outcomes. It might be more suitable for women with mild GDM. This meta-analysis also provides some significant benefits and risks of the use of metformin in GDM and help to inform further development of management guidelines.