Slow walking speed in elderly people predicts increased morbidity and mortality. We examined factors that may be associated with decreased habitual walking speed in older men and women. Older (range: 60-88 years, mean = 72.5 years) men (n = 25) and women (n = 24) were recruited. The Short Physical Performance Battery, body composition, VO(2peak) on a treadmill, VO(2) and rated perceived exertion during 10 minutes of walking at habitual gait speed and at a walking speed of 0.9 m/s, muscle strength, and level of physical activity were measured. VO(2peak) was strongly related to habitual gait speed (r = .744, p < .001) and remained significant even after controlling for age, muscle strength, and gender. Compared with the tertile of fastest walkers (mean gait speed, 1.37 +/- 0.04 m/s), the tertile of slowest walkers (0.87 +/- 0.02 m/s) were older (p < .001), shorter (p = .026), had lower lean body mass (p = .011), lower strength ( p < .001), less self-reported daily physical activity (p = .102), and higher relative (to VO(2peak)) intensity during walking at their habitual speed (65.3% +/- 3.9% vs 54.3% +/- 2.1% of VO(2peak), p = .013). VO(2peak) was strongly associated with habitual walking speed, suggesting that as aerobic capacity declines with age, the exertion associated with habitual gait speed increases. A slowing of walking speed may be a response to increased perception of exertion. The extent to which exercise training affects habitual gait speed and fatigue is not clear.