Predicting the financial distress of a business has been a significant subject among researchers and professionals for preventive measures against bankruptcy. Past studies have found that financial distress models’ accuracy can be very contextual and has proven to produce a lower accuracy rate in countries with low similarities than in the original place these models developed. In Indonesia, the conventional financial distress formula has proven to give a low accuracy result. On the other hand, new methods utilizing machine learning also has been introduced in financial distress cases in other countries and have proven to outperform the traditional statistical model. This study applies machine learning: Support Vector Machine (SVM) in predicting Indonesian financial distress cases. Using 30 financial ratios in 420 company’s financial distress cases in Indonesia in 10 years period, the SVM models produce 90% accuracy rate, far better than the Altman z-score formula (1968), Ohlson model (1980), Springate model (1974), and Zmijewski model (1984) applied to the same sample. These results indicate that the machine learning method is recommended to be applied in early financial distress prediction for business decision-making and risk mitigation.