Abstract: Existing models for portfolio optimization of commercial banks use the distribution of returns as a normal distribution, which does not conform to the characteristics of the actual yield, and most studies do not consider the impact of existing loans on yield and risk, making the portfolio risk assessment improperly. The portfolio optimization model considering the stock loan under the stable distribution can reflect the characteristics of the actual yield, correctly assess the loan portfolio risk, and control the bias to a certain extent, so that the loan portfolio obtains excess returns. Considering the requirements of risk dispersion, the introduction of risk concentration to restrict the allocation proportion of incremental loans of the loan portfolio to avoid the additional risks caused by a certain loan, thus establishing a new bank loan portfolio optimization model. By analyzing and optimizing the model characteristics, target function form and number of variables, SQP algorithm is selected to allocate the final incremental loan ratio. After testing with the actual data, the new model is simple and feasible, so it can be actually used in the selection of bank loans.

Key words: Stable Distribution,  , Stock Loan,  , SQP Algorithm,  , Risk Concentration