Abstract: In recent years, the influence of falling coal prices makes coal futures’ price volatility risk increases. The market environment leads the VaR model which based on the normal distribution could not measure coal futures price risk accurately. Therefore, how to measure coal futures price risk accurately becomes a serious problem. This paper based on the VaR model, tries using K-S tests to find other distributions, which can improve the risk metrics accuracy, to replace the normal distribution assumptions. The K-S test results show that coal futures returns comply with the hyperbolic distribution; probability density curves and Q-Q figures show that the hyperbolic distribution fits the reture series better than the normal distribution; VaR calculation and comparison shows that the hyperbolic VaRs are closer to history VaRs than the normal VaRs, and coking coal futures’ VaRs are greater than power coal futures’ VaRs. Therefore, the VaR model based on the hyperbolic distribution is more suitable for coal futures investors to measure risk; investment risk of coking coal futures is bigger than that of power coal futures.