The Log-ACD model which includes the "number of days" variable and Copula model are used respectively to fit the marginal and joint distribution of continuously rising and falling stock yield.The result of test shows that this model is comparatively better than conditional methods.In this paper,Shanghai stock 180 index is used for empirical analysis and conditional VaR is computed for risk analysis.The empirical result proves that this model is consistent to actual market,and investor can use "Figure of Rising and Falling Status" to analyze and estimate their act of investing.