<p>Numerical modeling has played a pivotal role in advancing inertial microfluidics, tracing its development from inception and offering deeper insights into the microscale phenomena governing inertial focusing. These computational approaches have simultaneously supported the proliferation of on-chip technologies. Initially adopted across diverse industries for passive and high-throughput operations such as trapping, separation, and sorting of particles, the greatest potential of inertial microfluidics lies in biomedical applications, where it serves as a cornerstone for processing cells in clinical and research settings. As the range of applications continues to expand, microfluidic devices are evolving into increasingly complex systems capable of handling diverse cell types and particles within miniature chip architectures. This growing complexity necessitates the enhancement of conventional numerical techniques and the integration of innovative computational approaches to address these emerging challenges. This review aims to provide an overview of the available numerical techniques, highlighting their advantages and limitations. We explore recent strides in computational inertial microfluidics, emphasizing advancements within the last four years and the emergence of innovative methodologies such as smoothed particle hydrodynamics. Furthermore, we describe the nascent role of machine learning in inertial microfluidics, noting its limited adoption compared to conventional microfluidics and highlighting the potential to transform the field, as well as challenges that need to be overcome.</p>