本文创造性地指出,股价波动具备先增长后下降的倒U形曲线特征,即库兹涅茨曲线特征,从而对最高值出现的时间点进行预测.本文首先对股价波动曲线进行了去除微小波动和类均线的光滑化处理,通过多项式函数回归和Logistic函数回归的拟合方法,来确定曲线增长部分的拐点,并给出寻找拐点和预测的方法.抽样检测结果表明该方法非常具有实效性. This paper creatively points out that the stock price volatility has the characteristics of inverted U-shaped curve which firstly increases and then decreases, that is, the Kuznets Curve, so as to predict the time when the highest value appears.Firstly, By means of polynomial regression and Logistic regression, the inflection point of the curve growth part is determined, and the method of finding the inflection point and the prediction is given. The results of sampling test show that the method is very effective.