The aim of this study is to determine the prevalence of and risk factors of dementia in the elderly with a low socio-economical status. The study is a cross-sectional and analytic study. Two hundred and four persons aged 65 years and above were selected by using the cluster sampling method. The presence of dementia was determined by Standardized Mini Mental State Examination (SMMSE). Twenty-three and 18 cut-off points were accepted as the existence of dementia in educated and uneducated older individuals, respectively. Chi-square test, Student's t-test, and logistic regression analysis were used for data analysis. The mean age of older subjects was 70.8+/-6.45 (S.D.); 71.5+/-5.9 in males and 70.3+/-6.8 in females (p=0.19). Ninety-one percent of older people had never attended school, 54.6% were living below the poverty level, and 97% of the residences were in squatter settlements. Dementia was determined in 46 older subjects (22.9%). Dementia was found significantly higher in females, in elderly subjects aged 80 years and above, in older subjects who were in a bad or very bad self-health status and in those who had a chronic disease, did not have any occupation, had a lower monthly income, and with a depressive score eight and above (p<0.05). In our study, the prevalence of dementia in older people was higher, in agreement with many reported studies. We assumed that this was due to the fact that the ratio of the uneducated, poor, and advanced aged subjects in this study was considerably higher. The different prevalence rates in studies demonstrate that the adoption of SMMSE for the study population, the validity-reliability analyses of tests and the standardization of diagnostic criteria are necessary.