Based on the 5-minute high frequency data from the Chinese stock market，and with the non-para-metric method， the realized jump volatility components ( the size，mean，standard deviation and arrival rate)are estimated，and the empirical results show that: 1) the realized jump volatility components can predict the excessive return of most of the 25 portfolios，with the linear and non-linear time series regression model; 2 )the realized jump volatility components have some explanation power for the portfolio return，with the linear cross Based on the 5-minute high frequency data from the Chinese stock market, and with the non-para-metric method, the realized jump volatility components ( the size, mean, standard deviation and arrival rate)are estimated, and the empirical results show that: 1) the realized jump volatility components can predict the excessive return of most of the 25 portfolios, with the linear and non-linear time series regression model; 2) the realized jump volatility components have some explanation power for the portfolio return, with the linear cross sectional regression model; 3) the realized jump volatility is possibly the drive force for the size effect and B / M ratio effect in the Fama-French 3-factor model.sectional regression model; 3) the realized jump volatility is possibly the drive force for the size effect and B / M ratio effect in the Fama-French 3-factor model．