Abstract: The fractured and vuggy reservoirs in deeply buried carbonate formation at SHB area Tarim basin are mainly deposited in fault systems ruptured by regional structural stress, the storage space consists of fractures caused by fissured faults and Karst vugs because of subsurface dissolution. Therefore, identification and evaluation of faults becomes the key step of the reservoir assessment. Considering the weakness of the seismic energy reflected from the deeply buried formation and the strong attenuation of the unconsolidated sand over the surface in the center of the deserts, the identification of faults confronts challenges with any one kind of seismic attribute. In this study, we developed a processing procedure for identification and characterization of faults with fusion of optimal seismic attributes' data set through understanding of the geological and geophysical characteristics in the target area. First, the fault caused diffusions in the seismic profiles have been identified as the suspected proofs of the existence of the faults or fractures. Secondly, methods of spectral whitening and dip angle directed filtering have been implemented for noise attenuation and frequency band extension. Next, the seismic data set after noise attenuation and frequency extension has been input into processing procedures of multiple seismic attributes calculation, such as coherence, ant tracking, and curvature. Moreover, identification and evaluation of faults in subsurface formation have been fulfilled with the fusion of multiple seismic attributes, and also the probable time window of the formation of the faults has been analyzed with the combination of the seismic profiles and attributes' slices. Finally, the sparse pulse inversion results constrained with the optimal seismic attributes and structural vectors have been applied to prediction of spatial distribution of fractures and vugs, and well consistency with the drilled results has been achieved.