Objective To test the reliable and valid of community screening interview for dementia-revised korean(CSID-K). Methods Study sample consisted of a dementia patient group and three control groups (depression, education, no-education), each of them was composed of 30 community residing elders in Korea. CSID-K was administered, and demographic characteristic, depression, and cognitive function were investigated. Results Split half reability (Guttmans'), internal consistency (Cronbach' s α), and inter-rater and test-retest (two weeks) reliabilities of the CSID-K were excellent. Scores on the CSID-K were significantly correlated with other measures for cognitive function. In addtion, scores on the CSID-K were significantly between the dementia patient group and three control groups. Conclusion CSID-K was a reliable and vaild instrument to screen dementia in Korean community residing old people. It could be used in clinical settings and for cross-cultural researches.