Osteoarthritis (OA) is the most prevalent degenerative joint disease, and acrylamide is a chemical produced when foods are processed at high temperatures. Recent epidemiological research linked acrylamide exposure from the diet and environment to a number of medical disorders. However, whether acrylamide exposure is associated with OA is still uncertain. This study was aimed at assessing the relationship between OA and hemoglobin adducts of acrylamide and its metabolite glycidamide (HbAA and HbGA). Data were taken from four cycles of the US NHANES database (2003-2004, 2005-2006, 2013-2014, 2015-2016). Individuals aged between 40 and 84 years who had complete information on arthritic status as well as HbAA and HbGA levels were eligible for inclusion. Univariate and multivariate logistic regression analysis s was performed to determine associations between study variables and OA. Restricted cubic splines (RCS) were used to examine non-linear associations between the acrylamide hemoglobin biomarkers and prevalent OA. A total of 5314 individuals were included and 954 (18%) had OA. After adjusting for relevant confounders, the highest quartiles (vs. lowest) of HbAA (adjusted odds ratio (aOR) = 0.87, 95% confidence interval (CI), 0.63-1.21), HbGA (aOR = 0.82, 95% CI, 0.60-1.12), HbAA + HbGA (aOR = 0.86, 95% CI, 0.63-1.19), and HbGA/HbAA (aOR = 0.88, 95% CI, 0.63--1.25) were not significantly associated with greater odds for OA. RCS analysis revealed that HbAA, HbGA, and HbAA + HbGA levels were non-linearly and inversely associated with OA (p for non-linearity < 0.001). However, the HbGA/HbAA ratio displayed a U-shaped relationship with prevalent OA. In conclusion, acrylamide hemoglobin biomarkers are non-linearly associated with prevalent OA in a general US population. These findings implicate ongoing public health concerns for widespread exposure to acrylamide. Further studies are still warranted to address the causality and biologic mechanisms underlying the association.