Aim There is a complex interrelationship between long sleep duration, frailty, chronic inflammation and mortality among the community‐dwelling middle‐aged and elderly population, which remains unclear and deserves to be investigated. The current study intended to explore these associations by using a prospective population‐based cohort study. Methods A total of 937 community‐dwelling middle‐aged and elderly people were enrolled. Sleep patterns of the study participants were categorized as short (<6 h), average or long (≥8 h). Sleep disturbance was defined by daytime dysfunction defined by the Pittsburg Sleep Quality Index. Frailty was defined as three or more phenotypes of Fried's Frailty. Results During an average of 4.7 years follow up, 72 (7.7%) study participants died. The adjusted hazard ratio (HR) for death of long sleepers was 2.42 (95% confidence interval [CI] 1.38–4.27), HR of long sleepers plus frailty was 2.37 (95% CI 1.35–4.19) and HR of long sleepers plus log interleukin‐6 was 2.11 (95% CI 1.19–3.76). Adjusted HR of daytime dysfunction was 1.79 (95% CI 1.09–2.94). The association between daytime dysfunction and mortality became statistical insignificant after further adjustment for frailty. Conclusions Long sleep duration, independent of frailty and interleukin‐6, was associated with 5‐year mortality in older adults. The relationship between daytime dysfunction and death diminished after adjusting for frailty. Geriatr Gerontol Int 2016; ••: ••–••.