To assess the effects of home massage rehabilitation therapy on the bed-ridden elderly. Alternatively allocated trial. Subjects' homes, three home nursing stations, 13 visit care stations and one day service centre in Aichi prefecture, Japan. Bed-ridden patients who were 65 years and above, no dementia, stable general condition, and receiving no rehabilitation therapy. Thirty-minute sessions of home massage rehabilitation therapy by a massage practitioner 2 or 3 days a week for three consecutive months or usual care. Barthel Index (BI), Subjective Satisfaction and Refreshment Scale, Apathy Scale and Self-rating Depression Score. Fifty-three subjects were recruited, 26 in the home massage rehabilitation group (HMG) and 27 in the routine care group without massage (RCG). The protocol was completed for 40 subjects, 22 in the HMG and 18 in the RCG. There were no significant differences between the baseline characteristics of both groups; age, presence of spouse, diseases associated with disabilities and use of day care rehabilitation (p = 0.76, 0.36, 0.94 and 0.71, respectively). The total BI score of the HMG (15.27+/-4.51) at baseline was nonsignificantly lower (p= 0.03) than those of the RCG (11.44+/-5.90). Subjective Satisfaction and Refreshment Scale, Apathy Scale and Self-rating Depression Score of both groups at baseline were matched (p = 0.12, 0.32 and 0.89, respectively). There were no statistical differences between the intergroup changes over time in BI, Subjective Satisfaction and Refreshment Scale, Apathy Scale and Self-rating Depression Score (p = 0.35, 0.08, 0.70 and 0.55, respectively). Home massage rehabilitation therapy did not show a positive effect on the bed-ridden elderly, either mentally or physically. We would require large-size trials to determine whether it is effective.