The kala-azar control program, adopted by the Fundação Nacional de Saúde-FNS (National Health Foundation) has not been able to reduce to an acceptable level the incidence of human cases. The diagnostic method utilized is a blood eluate immunofluorescence. A dogs diagnosed as infected is eliminated a mean of eighty days after the blood collection. The low sensitivity of the test used and the continuing residence of the infected dog in the region due to the elimination delay may be critical in the lack of success of this program. In this study, the FNS standard canine control method is compared to a strategy based on ELISA identification of infected dog and elimination within 7 days. In both study areas the canine seroprevalence was noted ten months before and ten months after the intervention. In the routine FNS area a 9% decrease in seroprevalence was noted, compared to statistically significant greater 27%, reduction (p = 0.0015) in the ELISA intervention area.