Thanks to the advancement of models and methodologies to handle large amounts of data in conjunction with the increased attention of national and international governments toward safety issues, this paper aims to provide a brief overview of the main models that are used to study pedestrian crashes. The reason why this type of analysis is conducted starts from three main research questions: What are the main datasets needed to study pedestrian accidents, what are the main models utilized, and what are the main gaps that emerge in the pedestrian safety field? This proposed state-of-the-art overview starts from the analysis of statistical approaches in the context of risk factor analysis to the most recent machine learning methods to evaluate pedestrian crash severity by emphasizing the purposes for which the models are used, why they are used, and the data needed to achieve the task. The results of the analysis show how the models could be classified and the main research gaps in this field that could be useful for researchers as starting points in their studies.