A number of factors more or less unique to multiple sclerosis have suggested that this disease may be particularly amenable to cell-based reparative therapies. The relatively focussed damage to oligodendrocytes and myelin at least in early disease implies that only a single population of cells need be replaced-and that the daunting problem of re-establishing connectivity does not apply. The presence of significant though partial spontaneous myelin repair in multiple sclerosis proves there to be no insurmountable barrier to remyelination intrinsic to the CNS: the therapeutic challenge becomes that of supplementing this spontaneous process, rather than creating repair de novo. Finally, the large body of available knowledge concerning the biology of oligodendrocytes, and the success of experimental myelin repair, have allowed cautious optimism that future prospects for such therapies are not unrealistic. Nonetheless, particular and significant problems are not hard to list: the occurrence of innumerable lesions scattered throughout the CNS, axon loss, astrocytosis, and a continuing inflammatory process, to name but a few. Here we review the progress and the areas where difficulties have yet to be resolved in efforts to develop remyelinating therapies for multiple sclerosis.