气测录井资料作为油气勘探开发活动中最及时的、最直接的，也是最重要的资料，具有获取井下地层信息及时、丰富，分析解释快捷的特点。对气测录井资料综合解释，是对录井成果认识水平的最终体现，也是地震、测井、井筒系列技术成果的集中反映，对油气勘探开发起基础作用。 气测录井校正方法包括非地层气异常数据的剔出、全烃背景值的校正、取芯钻井时全烃值的校正和钻时的影响校正方法。气测录井常规解释方法包括全烃曲线形态分析法、烃类气体比值图版法和色谱特征解释图版法等。论文在气测录井常规的解释方法和校正方法基础上利用储层内所含油气的成份特征和能量分布特征提出了一种多参数气测解释图版法，这种多参数解释方法可以更准确的找到油气层的分布情况。 建立以多参数为输入模式样本的人工神经网络解释模型，通过学习、记忆、自适应功能，提出一种有效的综合解释评价油气水层的新方法。该方法最大限度的综合运用多项原始资料，将气测录井参数量化及标准化处理，对储层特征进行描述，取得满意的效果。验证结果表明，应用人工神经网络技术对录井资料进行处理，能够有效提高录井资料解释准确度和油气层评价水平，并在此基础上初步开发了基于神经网络的气测录井资料解释软件。

Gas-logging data can aquair strata information timely, plentiful and analysis the data speedily. Comprehensive explaining the gas-logging data is the ultimate sign of logging results level, and also a concentrated reflection of the seismic, logging, wellbore series of technological achievements, based on oil and gas exploration and development. The correction metheods of gas-logging including picking the abnormal data, eliminating the total gas, reducing the total gas values by the influence of drilling time. The conventional interpretation methods of gas-logging including the analysis chart curve shape of total hydrocarbon, interpretation chart of gas ratio, etc. Researchers making a multi-parameter interpretation chart by using oil and gas component characteristic，this multiple parameters method can find the distribution of oil and gas layer more accurately. Establishing the interpretation model of artificial neural network which use multiple parameter input mode, through the learning and memory, adaptive function, and puts forward a effective new method of evaluation to explain oil and gas layer. The method using a number of sources maximumly, quantificating and standardizing gas logging parameters and describing reservoir characteristics, obtained satisfactory results. Results show that applied to process logging data by the artificial neural network technology can improve the interpretation accuracy level of logging data and evaluation of reservoir effectively. With the new method researchers also developing a gas-logging data interpretation software preliminarily.