The consequences of starting smoking by age 18 are significant. Early smoking initiation is associated with higher tobacco dependence, increased difficulty in smoking cessation and more negative health outcomes. The purpose of this study is to examine how closely smoking initiation in a well-defined population of American Indians (AI) resembles a group of Non-Hispanic white (NHW) populations born over an 80 year period. We obtained data on age of smoking initiation among 7,073 AIs who were members of 13 tribes in Arizona, Oklahoma and North and South Dakota from the 1988 Strong Heart Study (SHS) and the 2001 Strong Heart Family Study (SHFS) and 19,747 NHW participants in the 2003 National Health Interview Survey. The participants were born as early as 1904 and as late as 1985. We classified participants according to birth cohort by decade, sex, and for AIs, according to location. We estimated the cumulative incidence of smoking initiation by age 18 in each sex and birth cohort group in both AIs and NHWs and used Cox regression to estimate hazard ratios for the association of birth cohort, sex and region with the age at smoking initiation. We found that the cumulative incidence of smoking initiation by age 18 was higher in males than females in all SHS regions and in NHWs (p < 0.001). Our results show regional variation of age of initiation significant in the SHS (p < 0.001). Our data showed that not all AIs (in this sample) showed similar trends toward increased earlier smoking. For instance, Oklahoma SHS male participants born in the 1980s initiated smoking before age 18 less often than those born before 1920 by a ratio of 0.7. The results showed significant variation in age of initiation across sex, birth cohort, and location. Our preliminary analyses suggest that AI smoking trends are not uniform across region or gender but are likely shaped by local context. If tobacco prevention and control programs depend in part on addressing the origin of AI smoking it may be helpful to increase the awareness in regional differences.