Physical activity has been associated with a reduced risk of gestational diabetes mellitus, but inferences have been hampered by recall and selection bias. The authors examined the relation between recreational physical activity before and during pregnancy and risk of gestational diabetes mellitus in a prospective cohort study. In 1996-2000, 909 normotensive, nondiabetic women in Seattle and Tacoma, Washington, were questioned during early gestation about physical activity performed during the year before and 7 days prior to the interview during pregnancy. Compared with inactive women, women who participated in any physical activity during the year before experienced a 56% risk reduction (relative risk (RR) = 0.44, 95% confidence interval (CI): 0.21, 0.91). Women spending >/=4.2 hours/week engaged in physical activity experienced a 76% reduction in gestational diabetes mellitus risk (RR = 0.24, 95% CI: 0.10, 0.64), and those expending >/=21.1 metabolic equivalent-hours/week experienced a 74% reduction (RR = 0.26, 95% CI: 0.10, 0.65) compared with inactive women. Physical activity during pregnancy was also associated with reductions in gestational diabetes mellitus risk. Women who engaged in physical activity during both time periods experienced a 69% reduced risk (RR = 0.31, 95% CI: 0.12, 0.79). Findings suggest that efforts to increase maternal physical activity may contribute to substantial reductions in gestational diabetes mellitus risk.