The time trend of mesothelioma incidence and projections of future cases provide useful information for analyzing proposed public health interventions where asbestos exposure may be an issue, evaluating regulatory proposals, and estimating the remaining potential costs of programs to compensate individuals with asbestos-related diseases. We used the April 2008 release of Surveillance, Epidemiology, and End Results (SEER) data, which covers 1973 through 2005, to analyze the time trends in age-adjusted mesothelioma incidence and to estimate an age and birth-cohort model to project the number of future mesothelioma cases. The increase in the number of SEER cancer registries from 13 to 17 in 2000 had little effect on the time pattern of age-adjusted mesothelioma incidence, and the pattern over time of pleural mesothelioma was indistinguishable from the pattern for total mesothelioma defined as sum of pleural and peritoneal cases. Our analysis suggests that the SEER registries viewed as a sample of the U.S. population over-represents high mesothelioma incidence, a fact that we accounted for in our projections. For 2008 we estimate approximately 2,400 cases, with asbestos the likely cause in 58%. We project that asbestos will no longer be a factor in mesothelioma cases after the year 2042. For 2008 through 2042, we estimate slightly more than 68,000 total cases, with asbestos the likely cause in 34%.