Introduction: Research shows marked differences in the prevalence of dementia among different ethnic groups. However, there is a relative dearth of studies focusing on how ethnicity may influence dementia. The main aim of the current study is to test potential mediating effects of depression and level of education on the association between ethnicity and dementia. Methods: The sample for this study, consisting of 2,796 community-dwelling elderly people aged 60 years and older, was drawn from a cross-sectional national survey entitled ‘Mental Health and Quality of Life of Older Malaysians'. The Malaysian adapted version of the Geriatric Mental State-Automated Geriatric Examination for Computer-Assisted Taxonomy was used to assess dementia and depression. Data analysis was conducted using IBM SPSS version 19. Results: As expected, the result of the χ2 analysis revealed significant ethnic difference in the prevalence of dementia (χ2 = 58.05, p ≤ 0.001). The results of the mediational analyses revealed that depression (Z = 4.05, p ≤ 0.001) and education (Z = 2.78, p ≤ 0.01) significantly, but partially, mediate the association between ethnicity and dementia. This suggests that disparities of depressive symptoms and education contribute to ethnic difference in the prevalence of dementia. Conclusions: These findings provide more insight into how ethnicity may influence an individual's vulnerability to dementia. The limitations of the study along with implications of the findings and needs for further study are discussed.