Abstract: The main challenge of oil and gas exploration in Mahu sag can be summarized as "low, deep, thin and subtle", namely, low structure amplitude, deep target zone, thin layer and subtle reservoir. It is suggested that high resolution and amplitude fidelity of the seismic data are the key factors to solve these issues. This paper focused on model methods for slip sweep harmonic noise suppression, common offset frequency interference suppression, the surface and deep step inverseQfiltering technique, OVT (offset vector tile) domain processing, and applied them to Muhu sag. The real seismic data processing results show that the new technology can be used to effectively suppress the strong harmonic interference and random noise, improve the resolution of seismic data, widen the effective band of the high frequency of seismic signal end by about15Hz, which is helpful to characterize the Triassic unconformable pinchouts and identify the thin sand bodies; in the OVT domain, the processed prestack seismic gathers data are of high fidelity in amplitude changing with offset and azimuth information, and lay a solid foundation for high quality glutenite reservoir identification and prediction