Background: Cognitive impairment is a major public health concern in aging populations, and early identification of risk factors is critical. The red blood cell distribution width to albumin ratio (RAR) has emerged as a potential biomarker reflecting inflammatory and nutritional status, but its association with cognitive impairment remains unclear. Objective: This study investigates the relationship between RAR and cognitive impairment in older adults, and explores potential mediating variables that may influence this association. Methods: A total of 2,913 participants aged >= 60 years from the National Health and Nutrition Examination Survey (NHANES) 2011-2014 cycles were analyzed, including 1,291 with cognitive impairment. Logistic regression assessed the association between RAR and cognitive impairment, adjusting for potential confounders such as age, gender, race, education, marital status, weight, height, and comorbidities. Restricted cubic spline (RCS) analysis evaluated the dose-response relationship and identified nonlinear thresholds. Subgroup analyses explored interactions between RAR and demographic/clinical factors. Causal mediation analysis, using a generalized linear model with a probit link and adjusting for age, sex, race, and education, was performed to estimate total, direct, and indirect effects via bootstrap resampling. Results: RAR was positively associated with cognitive impairment (P < 0.05). RCS analysis revealed a nonlinear threshold, with RAR >= 3.2 significantly increasing the risk of cognitive impairment (OR = 1.24, 95% CI: 1.11-1.38, P < 0.001). Subgroup analysis showed significant interactions between RAR and cardiovascular disease (CVD), hypertension, and depression (P for interaction <0.05). Stratified analysis found a stronger association between RAR and cognitive impairment in individuals without hypertension, CVD, or depression. Mediation analysis indicated that CVD (P = 0.036) and depression (P = 0.032) partially mediated the relationship, with CVD explaining 4.49% of the total effect. Hypertension had no significant mediating effect. Conclusion: RAR is significantly associated with cognitive impairment, with a stronger association when RAR >= 3.2. CVD and depression partially mediate this relationship, suggesting RAR as a potential biomarker for cognitive impairment in older adults.