The reconstruction of the epidemic history of several HIV populations, by using methods that infer the population history from sampled gene sequence data, has revealed important subtype-specific and regional-specific differences in patterns of epidemic growth. Here, we employ Bayesian coalescent-based methods to compare the population history of the HIV-1 subtype B and F1 epidemics in Brazil from non-contemporary env and pol gene sequences. Our results suggest that after the introduction of the subtypes B and F1 into Brazilian population, around mid to late 1960s and late 1970s, respectively, these subtypes experienced an initial period of exponential growth with similar epidemic growth rates ( approximately 0.5-0.6year(-1)). Later, the spreading rate of both subtypes seems to have slowed-down since mid to late 1980s. This demographic pattern is very similar to that reported for the subtype B epidemics in high-income countries where HIV was initially transmitted through homosexual intercourse and injecting drug use, as in Brazil; suggesting that the characteristics of transmission networks may be a key determinant of the HIV epidemic growth pattern. It is important to note that most of the subtype B and F1 sequences used in this study come from the Southeast region that has been the most affected by the AIDS epidemic in Brazil, being responsible for around 63% of all AIDS cases reported since the early eighties; but may not represent the demographic trend of the HIV-1 epidemic in other Brazilian regions.