OBJECTIVE: The primary aim was to investigate the comparative effects of massage therapy and exercise therapy on patients with multiple sclerosis. The secondary aim was to investigate whether combination of both massage and exercise has an additive effect. DESIGN: Randomized controlled pilot trial with repeated measurements and blinded assessments. SETTING: Local Multiple Sclerosis Society. SUBJECTS: A total of 48 patients with multiple sclerosis were randomly assigned to four equal subgroups labelled as massage therapy, exercise therapy, combined massage-exercise therapy and control group. INTERVENTIONS: The treatment group received 15 sessions of supervised intervention for five weeks. The massage therapy group received a standard Swedish massage. The exercise therapy group was given a combined set of strength, stretch, endurance and balance exercises. Patients in the massage-exercise therapy received a combined set of massage and exercise treatments. Patients in the control group were asked to continue their standard medical care. MAIN MEASURES: Pain, fatigue, spasticity, balance, gait and quality of life were assessed before and after intervention. RESULTS: Massage therapy resulted in significantly larger improvement in pain reduction (mean change 2.75 points, P = 0.001), dynamic balance (mean change, 3.69 seconds, P = 0.009) and walking speed (mean change, 7.84 seconds, P = 0.007) than exercise therapy. Patients involved in the combined massage-exercise therapy showed significantly larger improvement in pain reduction than those in the exercise therapy (mean change, 1.67 points, P = 0.001). CONCLUSIONS: Massage therapy could be more effective than exercise therapy. Moreover, the combination of massage and exercise therapy may be a little more effective than exercise therapy alone.