OBJECTIVE: to explore the relationship between blood lipids/lipoproteins and cognitive function in the Chinese oldest-old. DESIGN: multivariate statistical analysis using cross-sectional data. SETTING: community-based setting in longevity areas in China. SUBJECTS: eight hundred and thirty-six subjects aged 80 and older were included in the sample. METHODS: plasma total cholesterol, triglycerides, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, blood pressure and fasting plasma glucose were measured and information about demographics and lifestyle was collected. Cognitive status was assessed using the Mini-Mental State Examination (MMSE). RESULTS: cumulative logit model analysis showed that triglyceride was significantly negatively associated with cognitive impairment. By general linear modelling, there was a significant linear trend of MMSE scores with the level of triglyceride, but not with levels of cholesterol after adjustment. The odds ratio (OR) of cognitive impairment (MMSE score < 18) was significantly reduced for the highest quartile of plasma triglyceride concentration (OR: 0.52, 95% CI: 0.33-0.84), but not for the second or third quartile, compared with the lowest quartile (adjusted models). There were no significant associations between cognitive impairment and cholesterol. CONCLUSION: we concluded that high normal plasma triglyceride was associated with preservation of cognitive function while lower concentrations were not in the Chinese oldest-old.