Pore structure and fracture determine reservoir quality in deep and ultra-deep tight sandstones, however, the limited availability of core data makes it challenging to describe the complexity of pore-throat structure and characterize the wide range of fractures. To fill this gap, a comprehensive analysis was conducted to evaluate pore structure and fractures from multiple perspectives. Integration of core, thin sections, scanning electron microscopy (SEM), high pressure mercury injection (HPMI) and nuclear magnetic resonance (NMR) tests are used to evaluate pore structure and characterize fracture of deep and ultra-deep tight sandstones in the Cretaceous Bashijiqike Formation in the Kuqa Depression. Pore structure and reservoir quality types can be divided into dominantly three types according to characteristics of NMR transverse relaxation time (T2) distribution and HPMI curves. Correlation analysis among HPMI and NMR parameters with petrophysical parameters indicate that maximum pore throat size (rmax), logarithmic mean of T2 (T2gm), irreducible water saturation (Swi), and reservoir quality index (RQI) are the most sensitive parameters for pore structure characterization. The Type Ⅰ pore structure is composed of intergranular and intragranular pores, and NMR T2 spectrum shows the highest magnitude, with the calculated mobile fluid saturation content also being the highest. The Type II and III pore structures are composed of intergranular pores, intragranular dissolved pores and intercrystalline micropores in clay minerals. T2gm of Type II and III are smaller, with the calculated mobile fluid saturation content being the lower.The fracture features and parameters including fracture density, porosity, aperture, and length are analyzed using image logs and core. The pore structure characteristics are quantificationally and qualitatively analyzed by NMR logs. Integration of pore structure and fractures by image logs and NMR logs are used for prediction of high quality reservoir and hydrocarbon productivity. High hydrocarbon productivity is obtained in well intervals with abundant natural fractures, and favorable pore structure will also contribute to high matrix porosity. Therefore, NMR logs and image logs are crucial for evaluating reservoir quality and predicting productivity when calibrated with core and analysis data. The comprehensive analysis integrating core data with geophysical well logs provide important insights for reservoir quality prediction of deep and ultra-deep tight sandstones.