With　the　development　of　microgrid　technology,　shortcircuit　fault　detection　and　location　have　become　key　challenges　for　the　reliable　operation　of　microgrids.　To　achieve　accurate　fault　location　in　AC　microgrids,　a　microgrid　simulation　model　was　constructed　in　MATLAB/Simulink　in　this　study.　The　short-circuit　faults　under　different　distributed　generation　capacities　were　analyzed.　Fault　features　were　extracted　using　wavelet　energy　spectrum　transformation,　and　a　BP　neural　network　was　employed　to　train　the　fault　location　model.　Simulation　results　indicate　that　the　proposed　method　can　achieve　early　fault　detection　and　precise　location,　particularly　demonstrating　high　accuracy　in　short-distance　fault　scenarios　under　grid-connected　mode.　©　2025　IEEE.