Inflammation has always been thought of as detrimental in the pathophysiology of multiple sclerosis (MS). However, emerging genetic data, magnetic-resonance-imaging studies, and immunopathological evidence challenge this simplistic view. The evidence leads to the conclusion that inflammation is tightly regulated, and that its net effect may be beneficial in MS, thus explaining some of the results from recent trials of anti-inflammatory agents. We argue that the use of anti-inflammatory drugs to treat MS may not be appropriate in all cases. Precise identification of the inflammatory pathways to be targeted in the different phases of the disease and the timing of such interventions are therefore crucial.