This paper primarily used statistical methods to construct a credit risk models on family controlled companies, discussed the agency problem on management and ownership are separated. That made it be possible to predict in advance the probability of family firms CEO power, for investors, financial institutes and stakeholders to estimate financial risks, so that financial risk losses can be minimized while corporate performance can be improved Empirical analysis, the author built up a credit risk model using K-S test, M-U test and logistic regression models, and to determine variables that can significantly affect corporate performance. Finding, multiple regression model it's risk sensitivity relatively high, in the case accounting data variables that can affect corporate performance in terms of the solvency and the operating performance. Liquidity ratio, times interest earned, fixed asset turnover and cash flow adequacy ratio that presented impact of significantly negative, while increasing them as early as possible to reduce the financial risk. Our empirical results suggest the CEO should increase cash flow adequacy ratio and adjust financing decision have to be enhanced the solvency, the operating performance and could reduce credit risk implementation.