Objective. The objective of this study is to systematically evaluate the effect of exercise on gestational diabetes (GDM). Methods. The databases of PubMed, Cochrane Library, Web of Science, Embase, CNKI, VIP, and Wanfang were searched to collect publications regarding physical exercises and GDM. The two researchers screened the literature, extracted the data, and analyzed the risk of bias of the included data using RevMan 5.3 software. The primary outcomes analyzed included the fasting blood glucose, 2-h postprandial blood glucose, glycosylated hemoglobin, premature delivery, cesarean section, neonatal macrosomia, premature rupture of membranes, and neonatal hypoglycemia. Results. A total of 9 studies with 1289 GDM patients were included. Compared with the control group, exercise could significantly reduce the 2-h postprandial blood glucose (
MD
=
−
0.62
, 95% CI (-0.91 to -0.34),
Z
=
4.29
,
P
<
0.0001
), improve HbA1c(
RR
=
−
0.47
, 95% CI (-0.81 to -0.13),
Z
=
2.69
,
P
=
0.007
), reduce the cesarean section rate (
RR
=
0.83
, 95% CI (0.71-0.98),
Z
=
2.25
,
P
=
0.02
), and decrease the incidence of neonatal macrosomia in GDM patients (
RR
=
0.57
, 95% CI (0.34-0.95),
Z
=
2.17
,
P
=
0.03
). Conclusion. Exercise intervention can improve the blood glucose level of GDM patients, such as 2-h postprandial blood glucose and HbA1c. Meanwhile, exercise can also reduce adverse pregnancy outcomes, such as premature birth and macrosomia. Therefore, prescribing exercise to GDM patients can effectively manage GDM and improve adverse pregnancy outcomes.