Cnidarians belong to the first phylum differentiating a nervous system, thus providing suitable model systems to trace the origins of neurogenesis. Indeed corals, sea anemones, jellyfish and hydra contract, swim and catch their food thanks to sophisticated nervous systems that share with bilaterians common neurophysiological mechanisms. However, cnidarian neuroanatomies are quite diverse, and reconstructing the urcnidarian nervous system is ambiguous. At least a series of characters recognized in all classes appear plesiomorphic: (1) the three cell types that build cnidarian nervous systems (sensory-motor cells, ganglionic neurons and mechanosensory cells called nematocytes or cnidocytes); (2) an organization of nerve nets and nerve rings [those working as annular central nervous system (CNS)]; (3) a neuronal conduction via neurotransmitters; (4) a larval anterior sensory organ required for metamorphosis; (5) a persisting neurogenesis in adulthood. By contrast, the origin of the larval and adult neural stem cells differs between hydrozoans and other cnidarians; the sensory organs (ocelli, lens-eyes, statocysts) are present in medusae but absent in anthozoans; the electrical neuroid conduction is restricted to hydrozoans. Evo-devo approaches might help reconstruct the neurogenic status of the last common cnidarian ancestor. In fact, recent genomic analyses show that if most components of the postsynaptic density predate metazoan origin, the bilaterian neurogenic gene families originated later, in basal metazoans or as eumetazoan novelties. Striking examples are the ParaHox Gsx, Pax, Six, COUP-TF and Twist-type regulators, which seemingly exert neurogenic functions in cnidarians, including eye differentiation, and support the view of a two-step process in the emergence of neurogenesis.