Background: Metformin is increasingly used in clinical practice for the treatment of gestational diabetes mellitus. However, its safety and long-term effects on fetuses exposed to metformin in uterus remain controversial., Methods: We systematically searched PubMed, Embase, and the Cochrane database (last search was updated on 1 May 2019) for randomized controlled trials comparing metformin with insulin. Two reviewers extracted the data and calculated pooled estimates by use of a random-effects model., Results: Twenty-four studies were included. Among these, seventeen RCTs ( N = 2828 participants) were included for quantitative analyses and seven studies were included only for qualitative synthesis. Metformin lowered the risk of pregnancy-induced hypertension ( p = .03; risk ratio (RR) = 0.64; confidence interval (95%CI) [0.44, 0.95]), large for gestational age babies ( p = .04; RR = 0.82; 95% CI [0.68, 0.99]), macrosomia ( p = .01; RR = 0.63; 95%CI [0.45, 0.90]), neonatal hypoglycemia ( p = .001; RR = 0.72; 95%CI [0.59, 0.88]), and neonatal intensive care unit admission ( p = .01; RR = 0.74; 95%CI [0.58, 0.94]). Metformin did not increase premature delivery ( p = .11; RR = 1.28; 95%CI [0.95, 1.73]), preeclampsia ( p = .45; RR = 0.89; 95%CI [0.65, 1.21]), caesarean delivery ( p = .20; RR = 0.94; 95%CI [0.85, 1.04]), small for gestational age babies ( p = .95; RR = 0.99; 95%CI [0.69, 1.42]). The long-term results seemed to have no adverse effect, but the information was still limited., Conclusions: According to our review, metformin may have potential benefits for pregnant women and newborns with no obvious adverse effects. However, even more studies are needed to provide evidence for the future use of metformin.