The importance of effective risk management has never been greater in rec ent decades. However, the estimate accuracy of risk measures, such as Value-at-Risk, remains a challenge. In this paper, we propose a novel non-parametric method to efficiently enhance the accuracy of risk estimation as our method can avoid model misspecification and fully explore the tail information contained in asset returns. The non-parametric Value-at-Risk measure is embedded into a risk hedge model to increase the ef