Levetiracetam is an anticonvulsant which is assumed to act by modulating neurotransmitter release via binding to the vesicle protein SV2A. This could have an impact on signalling in the pain pathway. The aim of this study was to test the analgesic effect of levetiracetam in central pain in multiple sclerosis. This was a randomized, double-blind, placebo-controlled, cross-over trial with levetiracetam 3000 mg/day versus placebo (6-week treatment periods). Patients with multiple sclerosis, symptoms and signs complying with central neuropathic pain and pain symptoms for more than 6 months, as well as pain intensity of more than 4 on a 0 to 10-point numeric rating scale were included in the study. The primary outcome measure was pain relief at the end of each treatment period as measured on a 6-point verbal scale. Eighty-nine patients were screened for participation and 30 patients entered the study. Twenty-seven patients were included in the data analysis. There were no differences in the ratings of pain relief (levetiracetam 2.4 vs. placebo 2.1, p = 0.169), total pain intensity (levetiracetam 5.3 vs. placebo 5.7, p = 0.147) or any of the other outcome measures (p = 0.086-0.715) in the total sample of patients. However, there was significant reduction of pain, increased pain relief and/or more favourable pain relief with levetiracetam than with placebo in patients with lancinating or without touch-evoked pain (p = 0.025-0.046). This study found no effect of the anticonvulsant levetiracetam in non-selected patients with central pain in multiple sclerosis, but an effect in subgroups with specific pain symptoms was indicated.