Background: The relationship between stroke subtypes and physical activity is unclear. Methods: Using data from 13,069 men and women aged 45–64 years who participated in the Atherosclerosis Risk in Communities Study, physical activity was assessed by self-report using the Baecke questionnaire at baseline (1987–1989). The American Heart Association’s ideal cardiovascular health guidelines served as a basis for the calculation of three physical activity categories: poor, intermediate, and ideal. Stroke and its subtypes were ascertained from physician review of medical records. Multivariable adjusted hazard ratios (HR) and 95% confidence intervals (CI) were calculated using Cox regression models. Results: During a median follow-up of 18.8 years, a total of 648 incident ischemic strokes occurred. Significant inverse associations were found between physical activity categories and total, total ischemic, and nonlacunar stroke in adjusted models (age, sex, race-center, education, cigarette-years). Compared with poor physical activity, the adjusted HR (95% CI) for ideal physical activity were 0.78 (0.62–0.97) for total, 0.76 (0.59–0.96) for total ischemic, 0.85 (0.51–1.40) for lacunar, 0.77 (0.47–1.27) for cardioembolic, and 0.71 (0.51–0.99) for nonlacunar stroke. Additional adjustments for waist-to-hip ratio, systolic blood pressure, antihypertensive medication, diabetes, left ventricular hypertrophy and laboratory parameters attenuated the HR. Further sex- and race-specific analyses revealed that the association was predominantly observed among males and among African-Americans. Conclusion: These data suggest a tendency toward a reduced risk of total, total ischemic, and nonlacunar stroke with higher levels of physical activity.