Abstract: The pre-stack anisotropic seismic inversion based on azimuthal seismic data is a key method for fracture detection and fluid identification of fractured reservoirs,which provides effective guidance for exploration and development of fractured reservoirs.However,the anisotropic inversion method under conventional constraints encounters challenges in terms of stability and reliability,with higher requirements being placed for fine characterization of exploration targets.Moreover,there exists a lack of direct inversion methods targeting the fracture density prediction and fluid identification of the fractured reservoirs with high-angle fractures in East China.Therefore,this paper first identified the sensitive factors for fluid identification using multi-well crossplots.Then,based on the poroelasticity theory and the linearized reflection coefficient equation for horizontal transverse isotropy(HTI) media,an azimuthal seismic reflection coefficient equation was deduced,which can reflect the variations of fluid factors and fracture density with the angle of incidence and azimuth angle.Finally,based on the advanced Markov Chain Monte Carlo(MCMC) algorithm,a two-step Bayesian anisotropic seismic inversion method with the L₁ norm constraint was proposed,achieving the fluid identification and accurate prediction of fracture density for reservoirs with high-angle fractures.The deduced reflection coefficient equation proved to be feasible through accuracy analysis.In addition,the method proved to be rational and reliable through synthetic seismogram testing and application in actual survey areas,offering a novel solution for the fracture density prediction and fluid identification in such reservoirs.