Schizophrenia is one of the most severe psychiatric disorders. Despite the knowledge accumulated over years, aetiology and pathophysiology remain uncertain. Research on families and twins suggests that genetic factors are largely responsible for the disease and implies specific genes as risk factors. Genetic epidemiology indicates a complex transmission mode, compatible with a multi-locus model, with single genes accounting for specific traits rather than for the entire phenotype. To better understand every single gene contribution to schizophrenia, the use of intermediate endophenotypes has been proposed. A straight communication between preclinical and clinical researchers could facilitate research on the association between genes and endophenotypes. Many behavioural tasks are available for humans and animals to measure endophenotypes. Here, firstly, we reviewed the most promising mouse behavioural tests modelling human behavioural tasks altered in schizophrenia. Secondly, we systematically reviewed animal models availability for a selection of candidate genes, derived from linkage and association studies. Thirdly, we systematically reviewed the studies which tested mutant mice in the above behavioural tasks. Results indicate a large mutant mice availability for schizophrenia candidate genes but they have been insufficiently tested in behavioural tasks. On the other hand, multivariate and translational approach should be implemented in several behavioural domains.