PredictingJiaying NiHung LeHien Thu Thi NguyenVan-Nam Huynh corporate financial status is one of the key research directions in financial risk management. Machine learning methods, on the other hand, have shown tremendous ability and potential in the domain of finance, particularly in financial distress prediction. The lack of access to financial data, the ambiguous definition of financial distress, and the differences in the methods used in the literature, however, have contributed to the confusion of researchers who are interested in the field. In this study, we revisit the concepts of financial distress and review recent papers talking about predictors, datasets, and models to construct early warning systems for financial distress prediction. We identify the current trends, present our comparative results of the main approaches, and discuss the advantages and disadvantages of each approach. While much effort was spent on predicting the likelihood of distress, we find that there are still some research opportunities. We conclude our study by discussing the need to develop explainable methods and the potential of using multiple sources of information. Through this work, we provide an overview of the field and categorize our understanding of common methodologies to tackle the challenges in distress prediction research.