Abstract: Thevolatility of coal price has brought great risk to coal enterprises and its downstream firms.Therefore,based on GARCH models,fluctuant features of coal is researched empirically,and we find that the price of coal has long memory and clustering in volatility,the coal market shows the characteristics of high-risk and high-return and the non-symmetrical effect that favorable news have a larger impact on the market than bad news.Then,the VaR of coal market is calculated by VaR-GARCH models,and by comparing with actual loss we find that VaR-EGARCH(1,1) model based on GED-distribution can describe the risk of coal market more effectively.