When investing in new stocks, it is difficult to predict the returns and risks of new stocks by general methods because of the lack of historical data.In order to deal with this problem better, two kinds of portfolio models with uncertain returns are proposed and solved by numerical experiments.Firstly, the portfolio models with linear transaction costs and nonlinear transaction costs are analyzed based on uncertainty theory.Then, in the numerical experiment, the numerical solution of the model is given and compared with the model without transaction cost.Finally, the relationship between risk and expected return is analyzed, which shows that the model with trans- action cost has better stability in numerical results.