Each of three hundred and ten canine serum samples from São Paulo, Brazil were submited for a Knott test, a commercial ELISA test (Witness, Merial) and indirect fluorescent antibody (IFA) test (Leidy) for diagnosis of heartworm disease. Seventy samples were positive; 19 on the Knott technique, 12 on ELISA and 61 on the IFA test. Fecal samples were submitted for 169 dogs. 102 were positive for several types of parasites, among them Ancylostoma caninum (99%), Trichuris vulpis (16.6%), Eimeria spp. (8.8%) and Toxocara canis (6.7%). IFA tests revealed anti-D. immitis antibody in dogs unassociated with intestinal parasite status. After serum absorption with antigenic extract of A.caninum and T.canis it was shown that only 61 out of the 79 anti-D.immitis antibodies positive samples were positive to IFA test. The specificity for ELISA and IFA test was 97.9% and 83.1%, respectively. The IFA test showed the highest false positive results. The accuracy of ELISA and IFA tests, in previously absorbed serum, was 93.8% and 81.9%, respectively. These findings demonstrated that both methods are reliable for the diagnosis of canine dirofilariasis, although cross reactions have ocurred.