Abstract: Objective To explore the predictive value of vitamin B2 (VB2), S-adenosylmethionine (SAM), and homocysteine (Hcy) for age-related cognitive decline in elderly patients (≥60 years old). Methods A total of 142 hospitalized elderly patients from Department of Geriatrics, the Affiliated Hospital of Xuzhou Medical University, between December 2023 and June 2024 were included. Their cognitive function was assessed through the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA). According to the presence of cognitive impairment, the patients were divided into two groups: an age-related cognitive impairment (ARCI) group and a healthy control (HC) group. Correlation analysis was conducted to explore the relationship between MMSE and MoCA scores, and folate metabolites. Furthermore, stratified multiple linear regression analysis was performed to assess the impact and predictive value of folate metabolites on cognitive function. Results Significant differences were observed between the ARCI group and the HC group in terms of age, education level, total bilirubin, direct bilirubin, total bile acids, TAU-181, Aβ1-42 levels, total albumin, MMSE, and MoCA scores (P<0.05). Correlation analysis showed that MMSE and MoCA scores were positively correlated with serum VB2 (r=0.354, 0.314; P<0.001), VB9 (r=0.345, 0.355; P<0.001), and SAM (r=0.424,0.399; P<0.001) levels, and negatively correlated with Hcy (r=-0.363, -0.325; P<0.001) levels. Stratified multiple linear regression analysis revealed that VB2, SAM, and Hcy were predictors of cognitive decline (P<0.01), explaining 45.9% of the variance in MMSE scores (F=9.05, P<0.001, R²=0.459) and 42.7% of the variance in MoCA scores (F=10.375, P<0.001, R²=0.427). Conclusions VB2, SAM, and Hcy have significant predictive value for age-related cognitive decline in the elderly population and can provide reference for early risk assessment.