The asymmetry of volatility could be correctly explained by jumps which also involve some information additionally.In order to improve the prediction of volatility,by employing the realized volatility and non-parametric jump detection method using high frequency data,this paper discusses the effect of jumps of different risk characteristics on future volatility based on the study of the asymmetry of volatility and conducts an empirical analysis with the SH indexes panel data from 2009 to 2014.The results indicate that systematic jumps of economic cyclical industry indexes bear significant effect on volatility prediction,which means a high degree of correlation between the market index and industry index; while those aperiodic industry indexes almost show no discernible leverage effect,with lower correlation of the market index.