Mediterranean type of visceral leishmaniasis (VL) is a zoonotic parasitic infection. Some provinces of Iran are endemic for VL while other parts are considered as sporadic areas. This study aimed to assess a combination of recombinant K26 and rK39 antigens as well as crude antigen (CA), derived from an Iranian strain of L. infantum, compared to direct agglutination test (DAT) for the detection of VL in humans and domestic dogs as animal reservoir hosts of the disease. A combination of rK26 and rK39 antigens and also CA was evaluated using indirect ELISA on serum samples of 171 VL confirmed humans (n = 84) and domestic dogs (n = 87) as well as 176 healthy humans (n = 86) and domestic dogs (n = 90). Moreover, 36 serum samples of humans (n = 20) and canines (n = 16) with other potentially infectious diseases were collected and tested for finding cross- reactivity. The results of ELISA were compared to DAT, currently considered as gold standard for the serodiagnosis of VL. The sensitivity and specificity, positive predictive and negative predictive values were calculated compared to DAT. The positive sera had previously shown a positive DAT titer ≥ 1:800 for humans and ≥ 1:80 for dogs. Analysis was done by MedCalc and SPSS softwares. Using the combination of rK26 and rK39 in ELISA, a sensitivity of 95.2% and a specificity of 93.0% % were found in human sera at a 1:800 (cut-off) titer when DAT-confirmed cases were compared with healthy controls; a sensitivity of 98.9% and specificity of 96.7%% were found at a 1:80 (cut-off) titer compared with DAT. A good degree of agreement was found between the combined rK39 and rK26-ELISA with DAT in human (0.882) and dog serum samples (0.955) by kappa analysis (p < 0.05). The ELISA using the CA test showed 75% sensitivity in human and 93.1% in dog serum samples as well as 53.5% specificity in human and 83.3% in dog,s sera, respectively. The combination of rK26 and rK39 recombinant antigen prepared from Iranian strain of Leishmania infantum showed high accuracy for the serodiagnosis of VL in human and domestic dogs. Further extended field trial with a larger sample size is recommended.