Visceral leishmaniasis (VL) is an important public health problem in the world, and control measures are insufficient to avoid the spread of this neglected disease. Dogs are important domestic reservoirs of Leishmania parasites in countries where VL is a zoonosis, representing a major source of infection between sand fly vectors and humans. In this context, a precise diagnosis of canine leishmaniasis (CanL) could help to reduce the number of human cases. Distinct approaches for the diagnosis of CanL have used recombinant proteins in serological assays. However, variable results of the antigens have been found, mainly to diagnosis asymptomatic cases. The present study used bioinformatics to select specific B-cell epitopes of four Leishmania infantum proteins, which had previously been proven to be antigenic in VL, aiming to produce a novel chimeric protein and to evaluate it for the diagnosis of CanL. Seven B-cell epitopes were identified and used to construct the chimera, which was analyzed in a recombinant format through an ELISA assay against a canine serological panel. A soluble Leishmania antigenic extract (SLA) was used as an antigen control. Results showed 100 % sensitivity and specificity for chimera, while when using SLA the values were 26.0 % and 96.4 %, respectively. The performance of chimera was compared with a commercial kit using asymptomatic and symptomatic dog sera, and the data showed that no false-negative result was found when the recombinant protein was used. However, when using the commercial kit, 40.0 % and 16.0 % of the false-negative results were found, respectively. In conclusion, the recombinant chimera showed an antigenic potential to be evaluated in new studies against a larger serological panel for the diagnosis of CanL.