Few studies have reported the relationship between smoking in middle age and long-term risk of impaired activities of daily living (ADL). We analyzed 2,276 men and women aged 47-59 years at the baseline survey of NIPPON DATA80 in 1980. At the follow-up survey in 1999, ADL was surveyed among 1890 survivors. Multivariate-adjusted odds ratio (AOR) and 95% CI of impaired ADL or of composite outcome of either death or impaired ADL according to baseline smoking status were calculated by multiple logistic regression analyses. In 1999, 386 participants were dead, and 75 participants had impaired ADL. Compared with nonsmokers, AOR (95% CI) of impaired ADL was significantly higher in current smokers at baseline (odds ratio [OR] 2.11 [1.09-4.06]). Compared with nonsmokers, AOR of impaired ADL was higher as the number of cigarettes increased (OR 2.04 [1.02-4.06] for <20 cigarettes/day and OR 2.35 [0.94-5.88] for >20 cigarettes/day; p for trend = .04). AOR of composite outcome for current smoking was 1.83 (1.37-2.41). Smoking in middle age would increase future risks of impaired ADL. Smoking cessation may be important to prevent future impairment of ADL as well as death.