The dissertation proposes three approaches for financial risk management. In the first topic, we investigate the stock return and risk of financial holding companies via Markov regime-switching model. The model reduces the disadvantage of traditional linear model, which disregard information of another regime if there exist structural change during the estimation periods. The empirical result shows that all financial holding companies have different stock risk between state 0 and state 1. Moreover, stock risks of all financial holding companies are significant lower after listing. That is, financial holding companies have diversification benefits after listing. In the second topic, we gauge the credit risk of guarantee issue in a bills finance company in Taiwan by a market-based model. Since bills finance companies engage in short-term loans, we renew the contract that can extend short-term loans to mid-and long-term loans. We find that the recovery rate, different industries and business cycle have significant impact on the credit risk of the bills finance company. In the third topic, we relax the assumption of Jarrow, Lando and Turnbull (1997), and propose an elaborate model to gauge the credit risk of Taiwanese bank loans. The empirical result indicates that the credit risk is heavily reliant on the recovery rate. Therefore, collateral value check procedure is very important, which has been found in previous topic. On the other hand, we find that the credit risk management is indifferent between banks participated in financial holding companies and others. That is, banks do not have better credit risk management if take part in financial holding companies. In conclusion, we expect approaches of the dissertation will be helpful for Taiwan¡¦s financial institutions to rise to the challenge of financial risk in the future.