Background The systemic immunity-inflammation index(SII) is a new indicator of composite inflammatory response. Inflammatory response is an important pathological process in stroke. Therefore, this study sought to investigate the association between SII and stroke. Methods We collected data on participants with SII and stroke from the 2015-2020 cycle of National Health and Nutrition Examination Survey (NHANES) for the cross-sectional investigation. Multivariate linear regression models were used to test the association between SII and stroke. Fitted smoothing curves and threshold effect analysis were applied to describe the nonlinear relationship. Results A total of 13,287 participants were included in our study, including 611 (4.598%) participants with stroke. In a multivariate linear regression analysis, we found a significant positive association between SII and stroke, and the odds ratio (OR) [95% CI] of SII associating with prevalence of stroke was [1.02 (1.01, 1.04)] (P < 0.01). In subgroup analysis and interaction experiments, we found that this positive relationship was not significantly correlated among different population settings such as age, gender, race, education level, smoking status, high blood pressure, diabetes and coronary heart disease (P for trend > 0.05). Moreover, we found an nonlinear relationship between SII and stroke with an inflection point of 740 (1,000 cells /µl) by using a two-segment linear regression model. Conclusions This study implies that increased SII levels are linked to stroke. To confirm our findings, more large-scale prospective investigations are needed.© 2024. The Author(s).