The proportion of Americans who smoke cigarettes has declined 50% since 1965. The effect on mortality of this considerable reduction has received little attention and is described in this study. U.S. national data were used to enumerate current, former, and never-smokers aged 35 years or older in 1987 and 2002. Mortality rate ratios were used to estimate smoking-attributable deaths among these groups, and corresponding age-adjusted smoking-attributable mortality rates (SAMRs) were calculated. There were 402,000 deaths attributable to smoking in 1987 and 322,000 in 2002. The SAMR for men aged 35 years or more was 556 deaths per 100,000 person-years in 1987, accounting for 24% of all male deaths. By 2002 the SAMR declined 41% to 329 and accounted for only 17% of deaths. The SAMR for women in 1987 was 175, accounting for 12% of deaths. By 2002 the SAMR among women had declined 30% to 122, representing 9% of deaths. The U.S. mortality rate attributable to smoking declined about 35% between 1987 and 2002. The impact of smoking on American society will diminish even further in the foreseeable future as smoking prevalence continues its decline among men and women.