OBJECTIVES: A dynamic multistate model was used to project the future burden of Chronic Obstructive Pulmonary Disease (COPD) in the Netherlands in relation to trends in demography and smoking. With the help of the model, different projections can be made that vary according to assumptions about future developments in smoking, about demography, and about COPD prevalence. METHODS: First, a simple prevalence projection of COPD was made, applying observed 1994 5-year prevalence rates on a projection of the population in 2010. That is, the prevalence projection uses prevalence of for instance the 55- to 59-year-old in 1994 and population projections to find prevalence of this same age group 10 years later. Second, a dynamic model that accounts for age and gender dependent incidence of COPD and trends in smoking prevalence was used. With the model we made incidence and smoking based projections, one assuming that smoking would remain at its 1994 levels and another using a scenario for future smoking based on observed trends in start and stop rates. The model basically uses prevalence of the 45-to 49-year-old in 1994 together with incidence and mortality to project prevalence of the 55- to 59-year-old 10 years later. RESULTS: The prevalence projections find an increase in COPD total prevalence of 27%. The model projections find a prevalence increase of 70% and 68%, with the lower assuming smoking at its 1994 levels. In this specific case, the prevalence projections find increases in COPD that are half those projected by the model. CONCLUSIONS: The results demonstrate differences between the extrapolation of prevalence rates using population projections and more complex projections with a dynamic model. These differences cannot solely be explained by trends in smoking prevalence: this is shown by the model projection with constant smoking prevalence. In this specific case, prevalence extrapolations underestimate future prevalence.