This a company is bankruptcy prediction method is needed by various parties, such as investors, accountants, governments, lenders, and management in order to predict the continuity of a company's operations in the future. Various bankruptcy prediction studies have been conducted to determine the most precise and accurate bankruptcy prediction model in predicting bankruptcy. This study aims to examine which of the Altman model, Zmijewski model, Springate model, CA-Score model, Grover model, and Ohlson model could significantly explain company bankruptcy and have the most accurate prediction accuracy. This research uses a descriptive quantitative approach. The sample is purposive sampling by taking 12 companies that were declared delisted from the Indonesia Stock Exchange in 2008-2022. The comparative sample is companies that are still listed on IDX with the same number and type, and randomly taken during the same period as the delisting company. The analysis technique in this research is binary logistic regression. The research results prove that of the six bankruptcy prediction models that can significantly explain company bankruptcy are the Zmijewski model, CA-Score model, and Ohlson model. However, the prediction model that has the most accurate level of prediction accuracy is the Zmijewski model. It is caused delisting companies that are the object of observation have a tendency of Earning After Taxes that obtained in a loss-profit or negative state and the amount of debt tends to be very large.. Keywords: delisting, bankruptcy prediction model, altman model, zmijewski model, springate model, ca-score model, grover model, ohlson model.