Aim To examine the efficacy of rehabilitation for elderly individuals with dementia at intermediate facilities between hospitals and home, based on the policies for elderly individuals to promote community‐based care at home and dehospitalization. Methods Participants were older adults with dementia newly admitted to intermediate facilities. A total of 158 in the intervention group who claimed Long‐Term Care Insurance for three consecutive months, and 54 in the control group were included in the analysis. The interventions were carried out in a tailor‐made manner to meet individual needs. The personal sessions were carried out three times a week for 3 months after admission by physical, occupational or speech therapists. Outcome measures were cognitive tests (Hasegawa Dementia Scale revised [HDS‐R] and Mini‐Mental State Examination), and observational assessments of dementia severity, activities of daily living (ADL), social activities, behavioral and psychological symptoms of dementia (BPSD) using a short version of the Dementia Disturbance Scale (DBD13), depressive mood, and vitality. Results Significant improvement in the intervention group was shown in cognitive function measured by HDS‐R (interaction F[1, 196] = 5.190, P = 0.024), observational evaluation of dementia severity (F[1,198] = 9.550, P = 0.002) and BPSD (DBD13; F[1,197] = 4.506, P = 0.035). Vitality, social activities, depressive mood and ADL were significantly improved only in the intervention group, although interaction was not significant. Conclusions Significant improvement by intervention was shown in multiple domains including cognitive function and BPSD. Cognitive decline and worsening of BPSD are predictors of care burden and hospitalization, thus intensive rehabilitation for dementia was beneficial for both individuals with dementia and their caregivers. Geriatr Gerontol Int 2013; ●●: ●●–●●.