Abstract: The complex mountain areas in Kuqa depression are characterized by complicated surface and underground conditions, large elevation difference, developed mountains and many gullies. The Paleogene gypsum salt rock suffered from serious deformation with large variations in thickness. The post-salt layer is high and steep and thrust imbricated faults are developed in the pre-salt target zone, resulting in low signal-to-noise ratio and poor image effect of seismic data. Over many years of research on seismic data processing in complex mountain areas, TTI anisotropic prestack depth migration technology has been developed for relief surfaces. The small smooth datum is used in prestack depth migration modeling. Thickness, time and velocity for static correction are calculated through microlog constrained tomography inversion and high-precision velocity model for shallow layers is established. A reasonable model is established for mid-deep layers by comprehensively using the data of surface outcrop, gravity-magnetic-electronic method, geology and drilling. The qualities and accuracies of the velocity models are improved by using TTI anisotropic parameter extraction and grid tomogrqaphic imaging technology. Based on the TTI anisotropic prestack depth migration processing in the complex mountain areas of Kuqa depression, the signal-to-noise ratio and imaging quality are significantly improved, which lays a foundation for regional geological study and trap identification