Financial ratios were used to build models of predicting corporate financial distress in the Taiwanese textile and construction industries. Further, in order to study the impact of sampling on the empirical results, the choice-based sample bias of the models were tested. The empirical results show that the quick ratio, debt ratio, working capital ratio, and ratio of operating profits were important variables influencing the financial distress of the textile industry, while the current ratio, debt ratio, asset turnover ratio, and earnings per share were important for the construction industry. Moreover, the test of choice-based sample bias shows that a sample with a lower proportion of financially distressed firms would result in a larger bias even if the percentage of correct classification would be increased.