Abstract: Based on three kinds of commonly used fluctuation measurement, this paper discusses the concrete form of multiple index fluctuant model. The multiple index fluctuant correlative estimation model is a kind of uniting forecast model, which is based on the multiplicative error model, with absolute daily returns, daily high-low range and daily realized fluctuation taken into consideration. After using the Shanghai stock markets data, the empirical results indicate that, there exist mutual effects between different methods to measure fluctuation, and the multiple index fluctuant model can remarkably increase the estimation and forecast accuracy of fluctuation.