中文摘要：

为了研究煤炭价格波动规律，以秦皇岛大同优混煤2003年3月-2017年5月现货价格为实证研究对象，分别采用普通最小二乘法和GARCH模型对煤炭价格时间序列进行拟合。实证结果表明：煤炭价格时间序列表现出随机波动趋势，GARCH（1，1）模型拟合效果优于最小二乘法；外部冲击会加剧煤炭价格波动，煤炭价格时间序列具备长期记忆性；波动率序列具备ARCH效应，ARCH项系数和GARCH项系数之和略大于1，说明煤炭价格波动的持续性较强。

英文摘要：

In order to study the law of coal price fluctuation in China, this paper selects Datong optimal mixed coal spot prices from March 2003 to May 2017 as the empirical research object, using ordinary least squares and GARCH model to fit the coal price time series respectively. The empirical results show that： the coal price series has a random fluctuation; GARCH （1,1） model is better than the least squares method in fitting coal price volatility; external impacts will aggravate the fluctuation of coal price; coal price series has long tern1 memory; the sum of ARCH coefficients and GARCH coefficients is slightly higher than 1, indicating a strong persistence of coal price fluctuation.