The massive emission of carbon from deforestation has brought this activity into being one of the most determinant of antropogenic global warming and climate change, as well as responsible for biodiversity loss in the tropical regions. In response to that, recent scientific work has focused on the development of fast, ample and precise estimations of forest biomass, leading to the methodological strengthening of future carbon-preserving projects. In the present work we investigate the relationship between satellite-based spectral vegetation indexes and land-based samples of forest biomass, seeking to develop a method of remote carbon estimation. Although the relation was significant, the spectral indexes only accounted for between 14% and 24% of the total carbon variation. This shows that in spite of probably being crucial for forest carbon estimation, vegetation indexes, who are intimately related to above-ground primary productivity (ANPP), do not predict carbon in an isolated manner, and a more precise model may need other variables (such as soil, topography, or humidity), or indexes capable of analysing further into the canopy.