Canine leishmaniasis (CanL) caused by Leishmania infantum (syn. L. chagasi, in Latin America), which is transmitted by the bite of phlebotomine sand flies, is endemic and affects millions of dogs in Europe, Asia, North Africa and South America. It is an emergent disease in North America. Early detection and treatment of infected animals may be critical in controlling the spread of the disease and is an essential part of human zoonotic visceral leishmaniasis control. The laboratory diagnosis of CanL still poses a challenge, despite progress made in the development of several direct and indirect methods. An effective diagnosis test, apart of being able to confirm a clinical suspicion in a single patient as well as to detect infection in asymptomatic dogs, should have high sensitivity, specificity and reproducibility; it must be simple, easy to perform, non-expensive, feasible in regional laboratories or adaptable for field conditions. Ideally, it should detect all Leishmania-infected dogs, preferentially using non-invasive collection of biological samples. In this paper we review the advantages and shortcomings of the available procedures for CanL diagnosis in the different phases, e.g. pre-patent and patent period of the infection and methods to determine the related immune response.