Serologic testing of wild birds for avian influenza virus (AIV) surveillance poses problems due to species differences and nonspecific inhibitors that may be present in sera of wild birds. Recently available competitive enzyme-linked immunosorbent assay (cELISA) kits offer a new species-independent approach. In this study we compare two commercial competitive cELISAs, using a total of 184 serum and plasma samples from 23 species of wild birds belonging to 10 orders. Thirteen samples were from experimentally high pathogenicity AI and low pathogenicity AI infected red-legged partridges (Alectoris rufa), 77 samples were from a flock of sentinel hybrid ducks confirmed infected by AI by real-time PCR, and 94 samples were from wild birds admitted to a rehabilitation center. Both ELISAs detected AI antibodies in the experimentally infected partridges, whereas hemagglutination inhibition (HI) was negative. Concordance in results between the two ELISAs was 51.5%. When specific subtype-H5/H7 HI-positive samples were considered for comparison, ELISA 1 appeared to perform better on ducks, whereas ELISA 2 appeared to perform better in other wild bird species. Overall, 68.2% of H5/H7 positive samples tested positive by ELISA 1 and 36% by ELISA 2. Both ELISAs detected AIV-antibody–positive samples negative by specific HI against 9 of the 16 existing hemagglutinin (HA) subtypes. Presumably this reflects either higher sensitivity of cELISA when compared to HI, presence of antibodies against HA subtypes not tested, or unspecific reactions. Performance of ELISA 1 on ducks appears to be comparable to in-house cELISA previously used by other authors in wild birds, but requires a relatively large sample volume. Alternatively, although ELISA 2 required a smaller sample volume, it was less effective at identifying HI-positive samples. The results reflect the necessity of validation of cELISA tests for individual species or at least families, as required by the OIE.