Adenosine is a ubiquitous neuromodulator, which carries out its biological task by stimulating four cell surface receptors (A(1), A(2A), A(2B), and A(3)). Adenosine receptors (ARs) are members of the superfamily of G protein-coupled receptors (GPCRs). Their discovery opened up new avenues for potential drug treatment of a variety of conditions such as asthma, neurodegenerative disorders, chronic inflammatory diseases, and many other physiopathological states that are believed to be associated with changes in adenosine levels. Knowledge of the 3D structure of ARs could be of great help in the task of understanding their function and in the rational design of specific ligands. However, since GPCRs are membrane-bound proteins, high-resolution structural characterization is still an extremely difficult task. For this reason, great importance has been placed on molecular modeling studies and, particularly in the last few years, on homology modeling (HM) techniques. The publication of the first high-resolution crystal structure for bovine rhodopsin (bRh), a GPCR superfamily member, provides the option of utilizing HM to generate 3D models based on detailed structural information. In this review we report, analyze, and compare the main experimental data, computational HM procedures and validation methods used for ARs, describing in detail the most successful results.