Objective Transplanted neural stem/precursor cells (NPCs) display peculiar therapeutic plasticity in vivo. Although the replacement of cells was first expected as the prime therapeutic mechanism of stem cells in regenerative medicine, it is now clear that transplanted NPCs simultaneously instruct several therapeutic mechanisms, among which replacement of cells might not necessarily prevail. A comprehensive understanding of the mechanism(s) by which NPCs exert their therapeutic plasticity is lacking. This study was designed as a preclinical approach to test the feasibility of human NPC transplantation in an outbreed nonhuman primate experimental autoimmune encephalomyelitis (EAE) model approximating the clinical and complex neuropathological situation of human multiple sclerosis (MS) more closely than EAE in the standard laboratory rodent. Methods We examined the safety and efficacy of the intravenous (IV) and intrathecal (IT) administration of human NPCs in common marmosets affected by human myelin oligodendrocyte glycoprotein 1‐125–induced EAE. Treatment commenced upon the occurrence of detectable brain lesions on a 4.7T spectrometer. Results EAE marmosets injected IV or IT with NPCs accumulated lower disability and displayed increased survival, as compared with sham‐treated controls. Transplanted NPCs persisted within the host central nervous system (CNS), but were also found in draining lymph nodes, for up to 3 months after transplantation and exhibited remarkable immune regulatory capacity in vitro. Interpretation Herein, we provide the first evidence that human CNS stem cells ameliorate EAE in nonhuman primates without overt side effects. Immune regulation (rather than neural differentiation) is suggested as the major putative mechanism by which NPCs ameliorate EAE in vivo. Our findings represent a critical step toward the clinical use of human NPCs in MS. Ann Neurol 2009;66:343–354