Behavioral changes in evolution have attracted the attention of many evolutionary biologists. Closely related species, or even individuals from different populations within a species, often exhibit remarkably different behaviors. Such behavioral diversification has been implicated as a cause of speciation in some cases, yet the mechanisms that produce and maintain these changes remain largely unknown. Drosophila melanogaster, an outstanding model organism with which to explore the causal link among the gene, neural circuitry, and behavior, provides an excellent entry point for a comparative approach to the origin of behavioral diversification with a single-gene and single-cell resolution. Here we survey studies that attempted to reveal the mechanistic bases for behavioral changes potentially associated with speciation, and we discuss the successfully identified candidate genes or neurons involved in such events. Although evolutionary developmental biology (evo-devo), a field devoted to the developmental and molecular basis of phenotypic diversity, has mostly been focused on morphological traits, the extension of this evo-devo approach to behavioral diversity will provide a comprehensive understanding of the genome-environment interactions underlying adaptive evolution.