IMPORTANCE: A better understanding of the role of atherosclerosis in the development of ischemic stroke subtypes could help to improve strategies for prevention of stroke worldwide. OBJECTIVE: To assess the role of carotid atherosclerosis in the association between major cardiovascular risk factors and ischemic stroke subtypes. DESIGN, SETTING, AND PARTICIPANTS: The prospective China Kadoorie Biobank cohort study was conducted in the general population of 5 urban and 5 rural areas in China, with a baseline survey obtained between June 2004 and July 2008. A random sample of 23 973 participants with no history of cardiovascular disease at enrollment who had carotid artery ultrasonographic measurements recorded at a resurvey from September 2013 to June 2014 were included. Data analysis was performed from July 1, 2016, to April 10, 2019. EXPOSURES: Major cardiovascular risk factors (ie, blood pressure [BP], blood lipid levels, smoking, and diabetes). MAIN OUTCOMES AND MEASURES: Carotid ultrasonographic measures of plaque burden (derived from number and maximum size of carotid artery plaques at resurvey) and first ischemic stroke during follow-up (n = 952), with subtyping (data release, August 2018) as lacunar (n = 263), probable large artery (n = 193), probable cardioembolic (n = 66), or unconfirmed (n = 430). Associations between cardiovascular risk factors, carotid plaque burden, and ischemic stroke subtypes were adjusted for age, sex, and geographic area. RESULTS: The 23 973 participants in the study had a mean (SD) age of 50.6 (10.0) years, and 14 833 (61.9%) were women. Systolic BP had a stronger association (odds ratio [OR] per SD, 1.51; 95% CI, 1.42-1.61) than plaque burden (OR per SD, 1.34; 95% CI, 1.26-1.44) with ischemic stroke, and the associations of systolic BP with each subtype of ischemic stroke were modestly attenuated by adjustment for plaque burden. After adjustment for BP, plaque burden had a stronger association with probable large artery stroke (OR, 1.43; 95% CI, 1.24-1.63) than with lacunar stroke (OR, 1.25; 95% CI, 1.10-1.43) but was not associated with probable cardioembolic stroke (OR, 1.06; 95% CI, 0.83-1.36). CONCLUSIONS AND RELEVANCE: Although BP was an important risk factor for all ischemic stroke subtypes, carotid atherosclerosis was an important risk factor only for large artery and lacunar strokes, suggesting that drug treatments targeting atherosclerosis may reduce the risk of stroke subtypes to different extents.