沪深大盘指数的收益率分布函数并不服从通常人们所认为的正态分布.因此,采用一种新的方法—非参数核密度估计,对沪深股指收益率分布进行拟合.该方法不仅很好地刻画了收益率分布的尖峰和肥尾特征,而且由此建立的VaR模型比一般的基于参数分布的VaR模型更能捕捉市场的风险特征,结论也更加准确. The distribution function of the returns of the Shanghai and Shenzhen stock market indexes does not obey the normal distribution that people usually think of. Therefore, a new method, non-parametric kernel density estimation, is used to fit the return distribution of Shanghai and Shenzhen stock indexes. It well characterizes the spikes and fat tails of the distribution of returns, and the VaR model established by this model can capture the market risk characteristics more than the general VaR model based on parameter distribution. The conclusion is also more accurate.