鄂尔多斯盆地西南部延长组储层是国内典型的致密砂岩储层之一，盆地中部延长组致密砂岩储层分类评价目前已引起了石油企业的重视。为了合理高效地开发鄂尔多斯盆地中部长8₁致密砂岩油藏，亟须开展储层分类，精准评价储层。以吴起地区长8₁致密砂岩储层为例，通过核磁共振实验数据计算目的层的分形维数值，并结合铸体薄片、扫描电镜等实验资料，对长8₁致密储层的微观孔隙结构进行定量研究；同时分析储层分形维数和储层物性的关系，运用基于核磁共振实验的分形理论对致密砂岩储层进行分类评价。结果表明：吴起地区长8₁致密砂岩储层岩石类型主要为岩屑长石砂岩，孔隙结构以粒间溶孔、长石溶孔为主；依据T₂谱曲线和分形维数法均将储层分为3类，各类储层对应相关的分形维数，分形维数与物性参数具有较好的相关性，以此表征了储层孔隙结构的复杂程度。将核磁共振实验与分形维数法结合定量地表征致密砂岩储层的微观特征，可提高储层分类的准确性。

The reservoir of Yanchang Formation in the southwest Ordos Basin is one of the typical tight sandstone reservoirs in China. The classification and evaluation of tight sandstone reservoir of Yanchang Formation in central Ordos Basin has attracted the attention of petroleum enterprises. In order to rationally and efficiently develop the Chang 8₁ tight sandstone reservoir in Ordos Basin, it is urgent to carry out reservoir classification and accurately evaluate the reservoir. Taking the Chang 8₁ tight sandstone reservoir in Wuqi area as an example, the fractal dimension of the target layer was calculated by NMR experimental data, and the micro-pore structure of the Chang 8₁ tight sandstone reservoir was quantitatively studied by combining the experimental data such as cast thin slice and scanning electron microscope. At the same time, the relationship between reservoir fractal dimension and reservoir physical properties was analyzed, and the fractal theory based on nuclear magnetic resonance experiment was used to classify and evaluate tight sandstone reservoirs. The results show that the rock type of Chang 8₁ tight sandstone reservoir in Wuqi area is mainly lithic feldspar sandstone, and the pore structure is mainly intergranular dissolved pores and feldspar dissolved pores. According to the T₂ spectral curve and fractal dimension method, reservoirs are divided into three categories. Each type of reservoir corresponds to the relevant fractal dimension, and the fractal dimension has a good correlation with the physical property parameters, which characterize the complexity of the reservoir pore structure. Combining nuclear magnetic resonance experiment with fractal dimension method to quantitatively characterize the microscopic characteristics of tight sandstone reservoir can improve the accuracy of reservoir classification.