Leishmaniasis is a significant public health concern, with dogs as the primary reservoir in urban scenarios and facilitating transmission. Diagnosing infected dogs is a crucial step for public health interventions, and the development of new diagnostic platforms can significantly enhance efforts in various regions worldwide. Given the limited availability of diagnostic methods in Colombia, this study evaluates the effectiveness of an Indirect Enzyme-Linked Immunosorbent Assay (ELISA) based on the recombinant protein rLicNTPDase-2 to detect Leishmania in infected dogs. Serum samples were collected from dogs in both endemic and non-endemic areas and classified as natural standards based on prior parasitological diagnoses. The results revealed 24 true positives (TP) and 9 true negatives (TN). Subsequently, the test was then validated with samples from symptomatic and asymptomatic animals, alongside the standards, yielding a specificity of 96 %, a sensitivity of 81 %, efficiency of 90.6 %, a positive predictive value of 92.8 %, and a negative predictive value of 89.6 %. The positive likelihood ratio (RV+) was 20, while the negative likelihood ratio (RV-) was 0.19, indicating high relevance and a robust clinical utility. The area under the curve (AUC) was 1.00, suggesting that the test has excellent discriminatory ability, significantly deviating from the reference diagonal. This is further supported by the significant difference(p < 0.0001) between TN and TP results determined by Fisher's exact test. Involving 163 animals showed 47 % positive and 46 % negative results with a significant difference (p < 0.05) in the mean optical density (OD) values between positive and negative samples. These findings indicate that the ELISA test effectively differentiates between positive and negative samples based on OD values. This study suggests that ELISA based on the recombinant antigen rLicNTPDase-2 could serve as a viable alternative for the serodiagnosis of leishmaniasis in canines in Colombia.