Abstract: In seismic profiles, low-grade faults are usually characterized by small dislocation, events distortion, weakened amplitude, short extension and small fault throw, and thus difficult to be identified and described. Based on the detailed analysis of observation system bin grid, S/N, frequency, velocity error and other factors that may affect the identification of low-grade faults during data acquisition and processing, we established a full wave field forward geological model with variance cube attributes and well data, and developed a targeted processing technique in this paper. Coherence cube attributes, curvature attributes along-layer slice, variance cube and other techniques are also used to effectively identify low-grade faults, and good results have been achieved.