Developing an effective and reliable power flow (PF) tool is of great significance for the steady-state analysis of autonomous AC/DC hybrid microgrids (MGs). In this paper, a generalized PF model is constructed based on holomorphic embedding (HE). With a well-designed embedding technique, the nonlinear PF problem is modeled into an embedded system, which comprehensively accounts for the distinct features of islanded hybrid MGs (such as droop-regulated distributed generations (DGs) and AC/DC interlinking converters (ICs), the variation of system frequency and its effect on admittance matrix and loads in AC part). The embedded system has a generalized structure with a constant sparse matrix, which not only enables it to be solved recursively and efficiently, but also facilitates wide applicability and convenient operation. Moreover, the proposed HE-based model features a deterministic germ that relates to a physical state, which allows it to inherit the merit of the canonical embedding in converging to upper-branch solution reliably and unambiguously, without dependency on initial estimates of state variables. The feasibility and applicability of the proposed model are validated on 3 test systems of different sizes. Furthermore, the computational efficiency and convergence performance are also evaluated. •Holomorphic embedding power flow modeling of islanded AC/DC hybrid microgrids.•Compatibility with droop regulation and distinct features of hybrid MGs.•Deriving reliable power flow solutions without dependency on initial guesses.•Generalized framework for wide applicability and efficient operation.•Experimental validation has been done on hybrid MGs of different sizes.