Background The association of circulating trimethylamine‐ N ‐oxide (TMAO) with stroke has received limited attention. To address this gap, we examined the associations of serial measures of plasma TMAO with incident ischemic stroke. Methods and Results We used a prospective cohort design with data pooled from 2 cohorts. The settings were the CHS (Cardiovascular Health Study), a cohort of older adults, and the MESA (Multi‐Ethnic Study of Atherosclerosis), both in the United States. We measured plasma concentrations of TMAO at baseline and again during the follow‐up using high‐performance liquid chromatography and mass spectrometry. We assessed the association of plasma TMAO with incident ischemic stroke using proportional hazards regression adjusted for risk factors. The combined cohorts included 11 785 participants without a history of stroke, on average 73 (CHS) and 62 (MESA) years old at baseline, including 60% (CHS) and 53% (MESA) women. We identified 1031 total incident ischemic strokes during a median 15‐year follow‐up in the combined cohorts. In multivariable analyses, TMAO was significantly associated with incident ischemic stroke risk (hazard ratios comparing a doubling of TMAO: 1.11 [1.03–1.18], P =0.004). The association was linear over the range of TMAO concentrations and appeared restricted to those without diagnosed coronary heart disease. An association with hemorrhagic stroke was not found. Conclusions Plasma TMAO levels are associated with incident ischemic stroke in a diverse population. Registration URL: https://www.clinicaltrials.gov . Unique identifier: NCT00005133.