To estimate the association between multiple domains of physical activity and risk of small-for-gestational-age (SGA) birth. We utilized data from 1,040 participants in the Latina Gestational Diabetes Mellitus Study, a prospective cohort of predominantly Puerto Rican prenatal care patients in Massachusetts. Physical activity was assessed by bilingual interviewers using a modified version of the Kaiser physical activity survey in early (mean = 15 weeks) and mid pregnancy (mean = 28 weeks). Physical activity (i.e., sports/exercise, household, occupational, and active living) in pre, early and mid pregnancy was categorized in quartiles. SGA was classified as <10th percentile of birth weight for gestational age. Pre- and early-pregnancy physical activity were not associated with SGA. In multivariable analyses, women with high total activity in mid-pregnancy had a decreased risk of SGA [risk ratio (RR) = 0.42; 95% confidence interval (CI) 0.21-0.82; p(trend) = 0.003] as compared to those with low total activity. Findings were similar for high household activity (RR = 0.69; 95% CI = 0.34-1.40; p(trend) = 0.26), active living (RR = 0.63; 95% CI = 0.35-1.13; p(trend) = 0.04), and occupational activity (RR = 0.79, 95% CI = 0.47-1.34; p(trend) = 0.26). High levels of sports/exercise were associated with an increased SGA risk without a significant dose-response association (RR = 2.14, 95% CI 1.04-4.39; p(trend) = 0.33). Results extend prior studies of physical activity and SGA to the Hispanic population.