This study compares five financial distress prediction models (Altman, Zmijewski, Grover, Springate, and Taffler) applied to Islamic banks in five ASEAN countries to identify the most conservative or lenient model and examine consistency of distress signals. Using secondary data from audited annual reports of Bank Muamalat Indonesia, Maybank Islamic Malaysia, Bangkok Islamic Bank Thailand, Al Amanah Islamic Investment Bank Philippines, and Islamic Bank of Brunei from 2015 to 2025, the study produced 55 bank year observations. Manual comparative analysis included descriptive classification, consistency index calculation, Wilcoxon signed rank test, Friedman test, and five robustness checks. Findings show Maybank Malaysia and Bangkok Islamic Bank Thailand are unanimously healthy. Al Amanah Philippines is unanimously distressed. Bank Muamalat Indonesia shows mixed signals with a consistency index of 0.24, indicating substantial model disagreement. Islamic Bank of Brunei shows borderline health. Altman is the most conservative model, followed by Springate, while Zmijewski, Grover, and Taffler form a lenient group with no significant differences. Model choice fundamentally alters distress classification for Indonesian and Bruneian Islamic banks. Regulators should use a composite or majority voting approach rather than relying on a single model. The sample is limited to one bank per country with no actual bankruptcy events, preventing direct validation of accuracy. This study contributes empirical evidence on model disagreement and offers practical guidance for regulators in the five ASEAN countries.