We aimed to investigate the association of recreational physical activity before pregnancy with gestational diabetes mellitus (GDM). A cross-sectional study was performed using self-reported data from the Pregnancy Risk Assessment and Monitoring System. The study population included 1,052 self-reported GDM cases and 10,351 non-GDM controls. Recreational physical activity in the 3 months before pregnancy was recalled in postpartum structured interviews. Compared to women exercising less than 1 day per week, women who exercised >5 days per week in the 3 months before pregnancy had a 31% lower odds of GDM (adjusted odds ratio [aOR]: 0.69; 95% confidence interval [CI]: 0.46, 1.03). Women who exercised 1-4 days per week had a 7% lower odds of GDM (aOR: 0.93; 95%CI: 0.72, 1.19). We did not observe statistically significant associations between pre-pregnancy physical activity and GDM risk. However, the magnitude and direction of association are consistent with previous observational studies. These studies collectively suggest a role of physical activity in GDM prevention.