ABSTRACT We evaluated a socially assistive robot (SAR) named Hyodol during a six-week intervention. This study enrolled 69 older adults with cognitive decline. To screen the eligibility, we have used the following three criteria, namely Korean-Mini-Mental Status Exam score ≤ 26). Clinical Dementia Rating 0.5–2), and Diagnostics and Statistical Manual V. Participants were divided into three groups based on their cognitive function (i.e. very mild cognitive impairment (vMCI), mild cognitive impairment (MCI), and moderate cognitive impairment (MOCI). The groups were instructed to use Hyodol at home for a six-week period. Baseline and post-intervention surveys were performed after six weeks to examine the changes in perceived health, depression, and cognitive function. The vMCI group showed a reduction in the depression score after the intervention (t = −2.447, p = 0.040), in comparison to their peers in the control group. Further, the MCI group showed an improvement in the cognitive function score after the intervention (t = 2.690, p = 0.021). No significant improvement was found among MOCI participants who used the SARs. The significance of this study was to examine whether participants with different levels of cognitive functioning would diverge after a period of intervention using the Hyodol SARs. Moreover, it presented preliminary data for services and policies for home care treatment targeted to cognitive decline in older adults.