An efficient analysis and optimization method is proposed to compensate the influence of asymmetric radome on an antenna by correcting amplitude and phase of the excitations. The asymmetrical and heteromorphic radomes are inevitable for the radar on high-speed aircraft. Many previous researches focused on the optimization of the radome structure and thickness to reduce the influence of radomes. However, the influence of complex streamlined radome cannot be compensated by merely optimizing the structure and thickness of the radome. Therefore, an alternative optimization method, optimizing amplitude and phase of feeds, is used in this paper. This paper adopts the active element pattern (AEP) technique, utilizing full-wave simulation method to extract the AEP for each antenna element and computing radiation patterns of array antenna by using vector composition of AEP. In combination with hybrid genetic algorithm-particle swarm optimization (HGAPSO), the antenna radiation characteristics can be obtained by updating excitations, which avoid the repeated full-wave simulation in the optimization process. Furthermore, the speed updating formula of PSO algorithm is improved combined with prior information, and the convergence speed is further increased. Finally, a 64 elements array antenna-radome system was optimized as an example in the cases of continuously adjustable phase and digital discrete phase.