ObjectivesPersonal Assistive RObot (PARO) interventions have been used to treat agitation and mood symptoms of dementia effectively. However, the effects of a PARO intervention on physiological and cognitive function are unclear. To examine the effects of a group-based PARO intervention for older adults with mild dementia. DesignUsing a group-based PARO intervention randomized controlled trial with 2-arm parallel groups. Setting and ParticipantsOlder adults with mild dementia aged 65 years or older from 4 dementia day care centers were recruited. MethodsPhysiological parameters were assessed using the finger tapping test (FTT) and heart rate variability (HRV). The Mini-Mental State Examination (MMSE), Geriatric Depression Scale–Short Form (GDS-SF), University of California Los Angeles loneliness scale–version 3 (UCLA-3), and Warwick-Edinburgh Mental Well-being Scale (WEMWBS) were assessed before the intervention, end of the intervention, and 1-month after the intervention. ResultsUsing a repeated-measures generalized linear model, significant time × group interactions were found in the MMSE [F(2, 115) = 19.54, P < .001], FTT [F(2, 115) = 4.87, P = .01], HRV high-frequency [F(2, 115) = 3.57, P = .03], and high-frequency/low-frequency ratio [F(2, 115) = 0.96, P = .01], UCLA-3 [F(2, 115) = 54.7, P < .001], GDS-SF [F(2, 115) = 3.36, P = .04], and WEMWBS [F(2, 115) = 5.93, P < .001]. Furthermore, psychological parameters improved significantly and continuously even 1 month after the PARO intervention was finished. Physiological parameters significantly improved at week 6, but the effects had diminished by week 10. Conclusions and ImplicationsA PARO intervention may effectively improve the physiological and psychological responses of people with mild dementia.