Previous research has yielded inconsistent results on the effects of exercise, smoking and alcohol use on cognitive impairment and dementia in old age. We analysed data from the Sydney Older Persons Study to see if these health habits were associated with cognitive functioning, dementia or Alzheimer's disease. Health habits were assessed in Wave 1 of the study, when the subjects were aged 75 years or over. Three years later, the subjects were tested for cognitive functioning and clinically examined for dementia and Alzheimer's disease. The analysis was restricted to the 327 subjects examined in Wave 2 who were non-demented in Wave 1. There were few significant associations between health habits and cognitive performance and these were not found consistently across cognitive measures. No associations were found with dementia or Alzheimer's disease. While these health habits do not affect risk for dementia and cognitive impairment in the very elderly, who are at highest risk for these disorders, we cannot discount a role at younger ages.