The high speed and manoeuvrability of cross-domain unmanned aerial vehicles (CDUAVs) significantly reduce the handover window between mid-terminal guidance stages, challenging reliable target acquisition. To address this, we propose an optimisation method for interceptor selection and field-of-view (FOV) cooperative coverage based on high-probability region (HPR) modelling. First, a predictive error covariance propagation model is constructed based on error propagation theory, and the HPR is established via eigenvalue decomposition. Second, the cooperative detection by heterogeneous interceptor seekers is formulated as a coverage optimisation problem with mixed FOV, and a cost minimisation model under complete coverage constraints is established. Finally, an improved genetic algorithm (IGA) is employed for solution, and a coverage area ratio screening mechanism based on two-dimensional close-packing theory is designed to enhance optimisation efficiency. Simulation results demonstrate that the probabilistic modelling approach for CDUAV HPR can accurately characterise the anisotropic distribution of target position uncertainty; the algorithmic mechanism reduces redundant computational load by 62.5% and shortens the optimisation time by 36.7%; the intelligent coverage optimisation framework provides a more generalisable solution for cooperative detection by heterogeneous interceptors’ seekers under conditions of target localisation uncertainty.