The induction of differentiation, as evidenced by benign growth characteristics, dendritic morphology, pigmentation capability, and a mature antigenic phenotype, is an attractive theoretical basis for therapy in human melanoma. Melanoma differentiation can be experimentally induced by modulating intracellular pathways involving protein kinase C, tyrosine kinases, and protein kinase A, or by modulating nuclear transcription with retinoids, DNA-damaging agents, and chemotherapeutic drugs. Other experimental differentiating agents include the amino acid tyrosine, histamine receptor antagonists, polyamine antagonists, dimethylsulphoxide, caffeic acid ester, and butyrate. The mechanisms involved in the actions of many of these agents are beginning to be understood and the pathways are often intersecting; cross-talk in the form of negative and positive feedback loops is extensive. Uncoupling of pathways is also seen, with some agents leading to simultaneous increases in both differentiated and transformed characteristics. While clinical benefits of this approach have so far been sparse, greater understanding of the cellular pathways of differentiation may open new therapeutic options in melanoma.