Abstract: This paper uses GRACHS model to carry out a dynamic modeling for finance time series, and based on this, puts forward the estimation method of VaR under skewed distribution. The empirical results of Shanghai copper futures markets suggest that, the benefits of Shanghai copper futures are time varying significantly and have obvious skewed distribution. The Kupiec test comparison of estimation for VaR of Shanghai copper futures shows that, the Estimation method of VaR based on the GARCHS model can improve the estimation accuracy for VaR under skewed distribution.