Background: Obesity is a potentially modifiable risk factor for cognitive impairment. However, sex-specific relationships between obesity and cognitive impairment in late life remain unclear.Objective: We aimed to assess sex differences in the association between various obesity parameters and cognitive impairment in a low-income elderly population in rural China.Methods: A population-based cross-sectional study was conducted to collect basic information from elderly residents aged 60 years and older from April 2014 to August 2014 in rural areas of Tianjin, China. Obesity parameters, including body mass index (BMI) and waist circumference (WC), and Mini Mental State Examination scores were measured, and the relationships between these variables were assessed.Results: A total of 1,081 residents with a mean age of 67.70 years were enrolled in this study. After adjusting for age, educational attainment, smoking status, drinking status, physical exercise participation, and the presence of diabetes and hyperlipidemia, blood pressure group; a high BMI was found to be associated with an increased prevalence of cognitive impairment in elderly women. Each 1-unit increase in BMI was associated with a 5.9% increase in the prevalence of cognitive impairment. WC was related to the prevalence of cognitive impairment in elderly men, and each 1-cm increase in WC was associated with a 4.0% decrease in the prevalence of cognitive impairment. However, there were no significant associations between WC and cognitive function in women or between BMI and cognitive impairment in men.Conclusion: A greater WC was positively associated with better cognitive function in low-income elderly men in rural China, whereas a higher BMI was associated with an increased risk of cognitive impairment in elderly women, independent of sociodemographic, lifestyle, and health-related comorbid factors. Our results suggest weight management of elderly women in rural China may have cognitive benefits. However, randomized controlled trials would be needed to confirm causality.