Reduced sleep duration has been increasingly reported to predict obesity. However, timing and regularity of sleep may also be important. In this study, the cross-sectional association between objectively measured sleep patterns and obesity was assessed in two large cohorts of older individuals.Wrist actigraphy was performed in 3053 men (mean age: 76.4 years) participating in the Osteoporotic Fractures in Men Study and 2985 women (mean age: 83.5 years) participating in the Study of Osteoporotic Fractures. Timing and regularity of sleep patterns were assessed across nights, as well as daytime napping.Greater night-to-night variability in sleep duration and daytime napping were associated with obesity independent of mean nocturnal sleep duration in both men and women. Each 1 h increase in the standard deviation of nocturnal sleep duration increased the odds of obesity 1.63-fold (95% confidence interval: 1.31-2.02) among men and 1.22-fold (95% confidence interval: 1.01-1.47) among women. Each 1 h increase in napping increased the odds of obesity 1.23-fold (95% confidence interval: 1.12-1.37) in men and 1.29-fold (95% confidence interval: 1.17-1.41) in women. In contrast, associations between later sleep timing and night-to-night variability in sleep timing with obesity were less consistent.In both older men and women, variability in nightly sleep duration and daytime napping were associated with obesity, independent of mean sleep duration. These findings suggest that characteristics of sleep beyond mean sleep duration may have a role in weight homeostasis, highlighting the complex relationship between sleep and metabolism.