Among the various diversified investment tools in the capital market, fund portfolio, with relatively low risks and good returns, is popular with investors. Traditional portfolio researches mainly consider the impact of market risk and ignore the coupled effect of credit risk and market risk, which often leads to underestimation of the overall risk of the portfolio. The edge distribution of the fund return is described with the GARCH model, and then the Copula function is used to describe the related structures between funds to establish the joint distribution model. Monte Carlo method is used to simulate the return sequence of funds in the fund portfolio. Finally, the value at risk of fund portfolios is calculated according to the loss function. The empirical results show that: ( i) the credit risk of the fund portfolio is not necessarily greater when its market risk is greater, and the fund portfolio can effectively diversify fund risk; ( ii) the CVaR of the coupled risk of fund portfolios is greater than the CVaR of market risk, which shows that coupled risk is better than market risk in measuring the overall risk when there is underestimation of the risk of fund portfolios; ( iii) compared with Gaussian model, the student t-Copula model can better depict the co-dependent structure of coupled risks.