Canine leishmaniasis (CanL), caused by Leishmania infantum, is widely distributed in many Mediterranean countries and is considered endemic in southern and central Italy with prevalence reaching up to 48.4%. Determination of the incidence would be useful as a measure of the risk of infection, then to evaluate the usefulness of control measures and to estimate whether a new focus is autochthonous or imported. This study was performed on two sites in the Apulia region of southern Italy, namely sites A and B. A total of 262 dogs were included in the evaluation of incidence, 94 farm dogs from site A and 168 dogs (92 farm and 76 kennel dogs) from site B. The incidence of infection was determined by using two different approaches: in site A by means of incidence density rate (IDR); in site B by the yearly seroconversion rate. In site A, the IDR was calculated at 4.25% dog-years; in site B the yearly incidence rate was of 9.52% (6.5% and 13.1% in farm and kennel dogs, respectively). The strength and weakness of the two different approaches (i.e. annual monitoring or monthly interval monitoring) for calculating the incidence of CanL in an endemic area have been discussed.