Objective: To determine an algorithm for molecular diagnosis of visceral leishmaniasis (VL) by kinetoplast DNA (kDNA) (RV1/ RV2) and internal transcriber spacer (ITS1) (LITSR/L5.8S) polymerase chain reaction (PCR), complemented by ITS1 PCR restriction fragment length polymorphism (RFLP), using peripheral blood or bone marrow aspirate from patients with suspected VL.Methods: Biological samples were submitted to the gold standard for the diagnosis of VL and molecular diagnosis represented by ITS1 PCR, kDNA PCR, and ITS1 PCR RFLP. The samples were obtained from seven groups: group I, 82 samples from patients with confirmed VL; group H , 16 samples from patients under treatment for VL; groupII, 14 samples from dogs with canine visceral leishmaniasis (CVL); group II, a pool of six experimentally infected sandflies (Lutzomya longipalpis); group IV, 18 samples from patients with confirmed tegumentary leishmaniasis (TL) and groups Ή and VI were from control groups without VII.Results: The following gold standard and molecular examination results were obtained for each of the seven groups: group I : parasitologic and immunochromatographic tests showed a sensitivity of 76.3% (61 of 80) and 68.8% (55 of 80), respectively, and a sensitivity of 97.6% (80 of 82) and 92.7% (76 of 82) by ITS1 and kDNA PCR, respectively. After ITS1 PCR RFLP (Hae III) analysis of the 80 positive samples, 52.5% (42 of 80) generated three fragments of 180, 70, and 50 bp, corresponding to the pattern of Leishmania infantum infantum; group Π : negative for the parasitologic methods and positive for IrK39 (100%, 16 of 16), presented 12.5% (2 of 16) of positivity by ITS1 PCR and 25.0% (4 of 16) by kDNA PCR; group III: positive in the parasitologic and serologic tests (100%, 14 of 14), presented 85.7%(12 of 14) of positivity by ITS1 PCR and kDNA PCR. ITS1 PCR RFLP showed that 83.3% (10 of 12) of the canine samples contained parasites with profiles similar to L. infantum; groupIVpresented amplifications by ITS1 PCR and kDNA PCR. ITS1 PCR products were analyzed by RFLP, generating a profile similar to that of L. infantum; group V: positive in the parasitologic examination (100%, 18 of 18), presented 72.2% (13 of 18) of the samples by ITS1 PCR positive. A total of 69.2% (9 of 13) showed profiles corresponding to a Viannia complex by ITS1 PCR RFLP; and group Ή and group W were negative by ITS1 and kDNA molecular tests. Comparing the molecular results with the parasitologic and serologic diagnosis from group I, almost perfect agreement was found ( κ both>0.80, P<0.001). ITS1 and RV1/RV2 PCR detected 90.2% (74 of 82) of the samples. Two samples positive by RV1/RV2 were negative by LITSR/L5.8S, and six samples positive by LITSR/L5.8S were negative by RV1/RV2. Therefore, these two systems complemented each other; they diagnosed 100% of the samples as belonging to the Leishmania genus.Conclusions: We suggest an algorithm for the molecular diagnosis of VL, which must consider previous parasitologic and serologic (immunochromatographic) diagnoses, and should combine kDNA and ITS1 to determine the Leishmania subgenus using RFLP as a complement method to define the L. infantum species.