Since the first studies on axonal regeneration, the optic nerve (ON) of higher vertebrates has been considered a good experimental system to investigate the failure of mature CNS neurons to re-grow after axotomy. The optic nerve is composed of a single population of fibers the RGC axons and, being separated from the rest of the brain, it is easily accessible to surgical manipulations. All the fibers can be transected without massive damage to the surrounding tissue, so their reaction to axotomy is not perturbed by extended inflammation processes. Another advantage of the system is the accessibility of RGCs. Being in the more internal retinal layer, RGCs are directly exposed to the humor vitreus, the liquid filling the posterior chamber of the eye. Pharmaceutical treatments are easily injected into the eye and, diffusing in the vitreus, can reach all the RGCs. Last but not least, functional recovery can be easily monitored in the optic nerve; measurement of electrical activity in response to visual stimuli in CNS regions that receive inputs from the retina such as superior colliculus or visual cortex allows evaluation of the re-growth of ON fibers and the restoration of connections. All the experiments carried out so far indicate that the failure of regeneration in the ON, as in the majority of the CNS districts, is a multi-factorial phenomenon, involving three classes of negative events. 1) RGCs die after axotomy: in the adult rat, their number is reduced to a very small percentage in a few weeks after the lesion. 2) The majority of mature axotomised RGCs are not programmed to re-start the process of axonal elongation that they displayed in immature stages. 3) The optic nerve environment contains molecules many of them upregulated after the lesion that are inhibitory for axonal growth. This review, focused on experiments performed in the mammalian optic nerve, traces attempts made to overcome each of these three obstacles, and maps progress towards a combined therapeutic strategy.