In a structure, a joint is considered as the weakest part, and it should not get separated when subjected to loading, so that an unstable collapse of structure can be avoided. It is important to investigate the failure in joint before it is used in a structure. Failure of a joint depends on various factors such as the geometry of joint configuration, sheet strength that are joined, rivet material used, cracks developed during joining, and many other. Self-Piercing riveting process is a new technology for joining sheet metals in automobile and aircraft industries. This process has many advantages over conventional joining processes. In this thesis, the failure of a self-piercing riveted joint is investigated. Failure of three different riveted configurations under 35m/s and 60m/s velocities were predicted using the general purpose non-linear finite element software LS-DYNA. This research is divided into three stages of work. In the first stage, a 2D simulation of riveting process is carried out over two Aluminum sheets. An r-adaptive methodology is utilized to acquire a higher accuracy of results and to avoid high element distortion. A parametrical study is then conducted to study the effect of rivet penetration velocity and adaptive mesh size varies the quality of the joint. In the second stage of work, a spring back analysis of joint is conducted to study the deformations of work piece after the riveting process. In the third stage, a Peel specimen, a U-shaped single riveted connection, and a U-shaped double riveted connection were investigated for failure under 35 m/s and 60 m/s velocities in both shear and tension testing conditions. Three different loading conditions were used for testing. The results from this study will show how process parameters can influence the quality of riveted joint, amount of deformations that occur in the work piece after the removal of rigid bodies, and failure load of SPR joint in different configurations.