The literature to date suggests a positive relationship between physiological indicators of physical fitness (such as aerobic capacity) and indices of cognitive performance. However, the complexity and cost of methods to measure physical fitness prohibit their use in large-population studies in cognitive aging research. In this study, a questionnaire measuring habitual physical activity was used as an indirect estimate of physical fitness, to predict performance in several cognitive domains in an age- and sex-stratified sample of 80 healthy older adults (55 years and older). Age effects were found on several measures of cognitive speed and fluency, but not on memory performance. Women were slower in sensorimotor speed than men but scored higher on memory tasks. No main effects of activity on cognitive measures were found, but two measures that assessed cognitive speed were sensitive to the age-by-activity interaction term. Subjective health also appeared to contribute to the explained variance in the same two indices of cognitive speed. Limitations of the use of activity questionnaires in cognitive aging research are discussed.