Botulinum toxin A has been suggested to be effective in the prophylactic treatment of migraine. However, only very few randomized, double-blind, placebo-controlled studies are available. We designed such a study with a specific focus on different injection sites. Sixty patients with a migraine according to the criteria of the International Headache Society were randomly assigned to receive either placebo in the frontal and neck muscles, or to receive 16 U botulinum toxin A in the frontal muscles and placebo in the neck muscles, or to receive in total 100 U botulinum toxin A in the frontal and neck muscles. The observation period was 3 months. In both treatment groups, 30% of patients showed a reduction of migraine frequency in month 3 by at least 50% compared with baseline, in the placebo group 25% of the patients showed such a reduction (P = 0.921). There were no significant differences between the three study groups with respect to reduction of migraine frequency, number of days with migraine, and the number of total single doses to treat a migraine attack. In the post hoc analysis, the reduction of all accompanying symptoms was significantly higher in the 16 U treatment group compared with the placebo group. In the 100 U treatment group significantly more adverse events occurred compared with the placebo group. All adverse events were mild and transient. Our study did not show any efficacy of botulinum toxin A in the prophylactic treatment of migraine. Only accompanying symptoms were significantly reduced in the 16 U but not in the 100 U treatment group. Future studies should focus on the efficacy of botulinum toxin A in specific subgroups of patients, on the efficacy of repetitive injections, and on other injection sites.