Objectives: To estimate smoking attributable mortality (SAM) in Taiwan for the years 2001 through 2020 under scenarios of reductions in smoking rates by 0%, 2%, 4%, and 10% per year. Method: The smoking attributable fraction (SAF) was used to calculate SAM from the risk experience in following up a large cohort ( 86 580 people) in Taiwan. Smoking rates were based on the 2001 National Health Interview Survey and other national surveys. An average 10 year lag was assumed between smoking rates and subsequent mortality. Results: In 2001, 18 803 deaths, or 1 out of 4 deaths ( 27%), in middle aged men ( 35 - 69 years old) were attributable to smoking. SAM has been increasing and will continue to increase if smoking rates remain constant or even if reduced annually by 2%. SAM would begin to decrease only if rates were to be reduced by at least 4% a year. Conclusions: The projected SAM in this study illustrates the seriousness of smoking caused mortality. Current efforts in tobacco control would lead to a progressive increase in SAM, unless efforts were doubled and smoking rates reduced by more than 4% a year. The urgency in requiring stronger tobacco control programmes to attenuate the staggering death tolls is compelling.