Abstract: In response to the need for enhanced cognitive detection performances of airborne early warning radars in complex geographic environments, as well as the problems of insufficient fidelity and low simulation efficiency in complex scenarios of traditional clutter modeling methods, a high-fidelity modeling method for inhomogeneous clutter based on multi-source prior information is proposed in this paper. A general mapping relationship between clutter blocks and multi-source prior information in the aircraft coordinate system is established by the proposed method, and the unified representation of multi-source prior information and high-precision modeling of inhomogeneous clutter are realized. On this basis, the fidelity of the simulated clutter is verified and analyzed in detail by combining measured clutter data, and a general fast simulation architecture for inhomogeneous clutter is designed based on the central processing unit and graphics processing unit heterogeneous platform. The experiment results show that the simulated clutter constructed by the proposed clutter modeling method has a good fit with the measured clutter, exhibiting high fidelity. The designed fast simulation architecture significantly improves the calculation efficiency of inhomogeneous clutter in complex large-scale scenarios, which has high engineering application value.