Adhesive joints have been used in several fields of engineering, and their applications are vast. Due to their easy and quick fabrication process, single-lap joints are a common configuration. The increase of strength, weight reduction and resistance to corrosion are some of the advantages of this kind of joint over traditional joining methods. However, stress concentrations at the overlap edges are one of the main disadvantages. There are very few accurate design techniques for the diversity of bonded joints that can be found in real applications, which constitutes an obstacle to the use of this bonding method in structural applications. This work aims at comparing different analytical and numerical methods in the strength prediction of single-lap joints with different overlap lengths (L O). The main objective is to evaluate which predictive method is the best. Adhesive joints were produced between aluminium adherends using a brittle epoxy adhesive (Araldite Õ AV138), a moderately ductile epoxy adhesive (Araldite Õ 2015) and a ductile polyurethane adhesive (Sikaforce Õ 7888). Different analytical methods were considered, together with two numerical techniques: cohesive zone models (CZM) and the extended finite element method (XFEM), allowing the comparative analysis. The analytical methods showed that they only give relatively accurate results in very specific conditions. The CZM analysis with the triangular law revealed to be a very accurate method, with the exception of joints with very ductile adhesives. On the other hand, the XFEM analysis was not adequate, especially for crack growth in mixed mode.