A cross-sectional study was carried out on bovine brucellosis in Addis Ababa dairy farms from November 2003 to April 2004. A total of 1,202 blood samples were collected from non-vaccinated, cross-bred dairy cattle. The Rose Bengal plate test (RBPT) was used as a screening test. Those serum samples reacting positively to RBPT were subjected to the complement fixation test (CFT) for confirmation. The RBPT detected 30 of 1,202 (2.5%) of the samples as brucellosis positive. The positive sera when further retested using CFT, 18 out of the 30 RBPT positive sera were confirmed to be positive. The prevalence of brucellosis based on CFT in the study area was 1.5%, and all positive sera were from female cattle. Result of the questionnaire survey revealed that percentage of 4.4% abortion and 9.5% retained fetal membranes. Abortion and retained fetal membranes were associated with Brucella antibodies (P < 0.05). A total of 153 cattle attendants and owners in the farms were interviewed, and 73.5% were found to have no knowledge of brucellosis, only 20.8% wear protective gloves during handling aborted material and 39.6% responded that they consume raw milk. Results of this study showed that prevalence of bovine brucellosis in the study area is low and a test-and-slaughter policy can be used in order to control the diseases in dairy farms of Addis Ababa.