This study explores the present forest management systems in place and their impact on the genetic diversity and the ecology of Hymenaea courbaril in the Brazilian Amazon forest, specifically in the Tapaj6s National Forest. As part of the Dendrogene Project (EMBRAPAIDFID), this study has as its main objective to evaluate the effects of Reduced Impact Logging on the species' genetic variability and ecological processes and to predict the likely impacts in the short- and long-term. The specific objectives of this thesis are to verify the quality of the. forest inventory used in RIL at Tapaj6s National Forest, to determine impacts of harvesting on forest structure and the species' reproductive population and spatial pattern, to determine Hymenaea . courbarifs genetic diversity, spatial genetic structure, pollen flow and mating system, and assess the impact of logging on these processes, and finally to infer on the impact of logging on Hymenaea courbarifs sustainability in the long-term using the modelling software Eco-gene. The results demonstrate the critical importance of correct botanical identification for the sustainability of forest management. The assessed forest inventory was highly inaccurate in relation to botanical identification and severely underestimated trees diversity which might compromise the capacity of accurately analysing species' spatial distribution and forest timber stock. The results also show that current forest management practices have negative impacts on the H. courbaril species including a reduction in reproductive population, spatial distribution and demography. The results of logging scenarios run through Eco-gene suggest that the harvesting cycle for Hymenaea courbaril should be between 120 and 150 years in order to reach pre-logging levels. More importantly, the model suggests that current logging practice strongly affect the species' population both ecologically and genetically in the long-term. These impacts were also observed for scenarios up to 90 years unless logging intensity is very low (10 - 30%). Moreover, the model results indicate that the current practise is not economically viable as the volume to be logged in future harvests will be acutely reduced. The results of this thesis provide forest planning and management with specific scientific criteria from a heavily logged species which can help to more effectively manage not only H. courbaril, but can also be used to move toward the sustainability of other tree species in the Amazon forest.