Optimization of genetic diversity management in terms of conservation and breeding purposes relies on an in-depth understanding of the genetic factors and evolutionary forces acting on the variability of the traits of adaptive and agronomic interest. In synergy with the quantitative and molecular genetic approaches, evolutionary approaches based on the analysis of patterns of nucleotide diversity allows identifying genomic regions harboring molecular signatures of domestication and adaptation. Through the definition of the most likely history of cultivated sorghum based on sequence diversity information from 20 accessions representatives of the sorghum worldwide diversity, demographic events which have affected the evolution of sorghum genes are taken into account to identify non neutral genes (i.e. domestication or adaptive genes). The domestication model developed, based on 200 000 polymorphisms identified on around 15 000 genes, allowed to estimate a domestication event around 5600 years BP, this result being in accordance with previous estimations. Bottleneck strength, importance of cultivated population expansion and bidirectional migration rate were also evaluated and allowed to reconstruct sorghum history. This information was then used to identify genes potentially involved in the domestication process or of adaptive importance. As an example, the analysis of the differentiation between the cultivated and wild pools allowed the identification of 64 genes harboring non neutral diversity patterns. It is interesting to note that a significant part of them is involved in cellular growth control, responses to nutrient availability, growth factors or abiotic and biotic stress.