在煤田勘探开发过程中，煤工业分析是一项极其重要的工作，它不仅是了解煤质特性的重要指标，同时也可以作为评价煤质的一个标准，对于初步判断煤的性质、种类、各种煤的加工利用效果和其工业用途有着重大的意义。 煤工业分析通常可由煤样实验室分析、测井体积模型法和概率模型法来确定。过去，煤工业分析主要是由煤样实验室分析获得，但这种方法有着不可避免的一些缺陷。然而，测井资料中不但包含着丰富的原始状态下地层的物理信息，而且作为其测定成果的测井曲线能够携带大量的和连续的地层地质信息。相对于煤工业分析的实验测试法，地球物理测井方法具有不扰动被测煤样的自然结构和原始状态等优点，很好的弥补了实验室测试的不连续等缺陷。 论文以HLC勘查区实际资料为依据，以数理统计、回归分析方法为手段，以预测勘查区5#煤工业分析指标为研究目的，开展了以下内容的研究： 首先，分析了煤的工业分析指标与测井幅值参数之间的内在联系，论证了利用测井幅值参数预测煤工业分析指标的可行性，为后续二者之间的定量研究奠定了理论基础；其次，分别建立煤的工业分析指标与测井曲线幅值参数的一元回归预测模型和多元逐步回归预测模型；最后，对这些所建立的预测模型的预测效果进行比较，选取模型显著性最高、拟合度最好、回归残差最小的回归模型作为研究区的最佳预测模型，并对选取的预测模型的应用效果进行了检验，取得了理想的效果。该方法在研究区取得的良好效果表明：在煤田勘探开发过程中，利用测井资料来预测煤工业分析指标是一种快捷、经济、实用的手段，可以推广和应用。

Proximate analysis of coal is a very important work in the process of coal exploration and development. That is because the proximate analysis of coal not only is one of the important indicators to understand the characteristics of coal, but also can serve as a standard of evaluation of coal quality, being of great significance for preliminary judgging the coal peculiarity, type, all kinds of coal processing effect and its industrial application. The proximate analysis of coal usually can be determined by the methods of coal sample laboratory test, logging volume model and probability model. In the past, the proximate analysis of coal is mainly obtained by coal sample laboratory analysis, but this method has some inevitable defects. However, in the logging data not only contain rich formation under the primitive state of physical information, also carry plenty of continuous formation and geological information as the determination results of logging curve. Relative to proximate analysis of coal of experimental testing method, the geophysical well logging method is simple, need not disturbing the natural structure of coal sample and the state of nature, it makes up for the lab test of those defects well. The paper based on actual data of HLC exploration area, taking 5# coal industrial analysis index prediction of exploration area as research purpose By means of mathematical statistics, regression analysis method, research was carried out by the following: Firstly, analyzed the inner link between the index parameters of proximate analysis of coal and amplitude parameters of well logging curves, demonstrating the use of log amplitude parameters to predict the feasibility of the coal industrial analysis index, laid the theory groundwork for the research behind; secondly, we set up the monadic regression prediction models and multivariate stepwise regression prediction models between the proximate analysis of coal and amplitude parameters of well logging curves; finally, by comparing the prediction effect of prediction models established anterior, then choosing the regression model which model significance is highest , fitting degree is greatest and regression residual is minimum as the best prediction model of research area.And the precision of prediction model was inspected, eventually obtained a ideal effect. The good results obtained in the study area showed that in the process of coal field exploration and development, using logging data to predict proximate analysis of coal is a fast, economic and practical means， should promotion and application.