According to the complexity and uncertainty in financial market，in order to more flexibly measure the volatility risk of financial market，this paper presents a class of semiparametric CAViaR model by using the nonlinear influence of former term fitting information of Fourier series on current VaR risk value to select suitable priori distribution，the related likelihood function is constructed based on asymmetric Laplace distribution，the posterior distribution of the parameters is deduced，and thus，the Bayesian deduction for CAViaR model is implemented.Finally，the feature of the risk volatility of Shanghai Composite Index is studied based on Bayesian CAViaR model and its results show the risk volatility of Shanghai Composite Index has autoregression.