In recent decades, Lagrangian methods such as Smoothed Particle Hydrodynamics (SPH) have gained popularity for predicting multi-phase problems, such as atomization, requiring up to billions of discretization points [1]. In contrast to grid-based methods, in SPH, the numerical treatment of the phase interface is inexpensive, and high parallel computation efficiencies are achievable. However, particle methods are still comparatively young and not fully mature. Consequently, in the past years efforts have been made to further improve the suitability of the SPH method to conduct atomization simulations. In this study, we employ a state-of-the-art SPH code using adaptive spatial resolution to further reduce the computational costs. The proposed approach is demonstrated using a well-studied generic prefilming airblast atomizer, which requires resolving different length scales ranging from a few micrometers to several centimeters. Hence, a multi-resolution approach is indispensable. It is well known that primary atomization is a highly dynamic and chaotic process. In most cases, the droplet statistics generated by the atomizer, such as size and velocity distributions, are of primary interest. In Figure 1 a single breakup event as predicted by this method is shown. As a quantitative result of this study, the atomization products predicted with adaptively refined and uniform particle distribution are compared. These comparisons aim to demonstrate, that the multi-resolution method is capable of accurately capturing the primary atomization characteristics.