Insulin resistance (IR) is an independent risk factor of cardiovascular disease. Recent research has proposed a new inexpensive and reliable indicator of IR: triglyceride glucose index (TyG index). We aim to evaluate the dose-response association between the TyG index and stroke through meta-analysis. Literature published from inception until October 2021 were searched in PubMed, Embase, Cochrane Library and Web of science. Cohort studies published in English and focusing on the association between the TyG index and stroke were included in our meta-analysis. I-2 statistic and Chi-square were used to assess the heterogeneity. When I-2 >= 30 % or p <= 0.10, the random-effect model was used to pool the effect; otherwise, we chose the fixed-effect model. Eleven cohort studies, including 5 721 077 subjects and 95 490 stroke patients, were included in our study. After pooling the effect adjusted by multiple confounders, we found that compared with the lowest baseline TyG index group, the highest one was independently associated with increased stroke risk (RR: 1.27; 95 % CI, 1.24 to 1.29; I-2 = 6 %). Dose-response meta-analysis showed that the association between the two variables had a non-linear trend (p-nonlinearity < 0.0001; p-heterogeneity = 0.28). Subgroup analysis showed that the risk of ischemic stroke was positively correlated with TyG index (RR: 1.48; 95 % CI, 1.34 to 1.62; I-2 = 15 %), while we did not observe this correlation in hemorrhagic stroke patients. In patients with type 2 diabetes and acute coronary syndrome, the TyG index was linearly correlated with incident strokes. In conclusion, elevated TyG index is the independent risk factor for incident strokes (especially ischemic stroke).