To　detect　and　locate　the　short-circuit　fault　area　quickly　and　accurately　in　the　microgrid　can　reduce　its　impact　on　the　large　power　grid　under　grid-connected　mode.　Therefore,　it　is　necessary　to　study　the　waveform　characteristics　of　short-circuit　fault　current　in　the　microgrid　under　grid-connected　mode.　In　this　paper,　a　short-circuit　fault　simulation　model　is　established　in　PSCAD.　And　the　current　information　is　monitored　at　the　connection　points　between　feeders　and　common　bus　of　microgrid.　Wavelet　energy　spectrum　algorithm　is　used　to　analyze　the　detected　current　waveforms,　and　the　characteristics　of　current　waveforms　are　studied　when　various　types　of　short-circuit　faults　occur.　The　simulation　results　show　that　the　high-scale　wavelet　energy　spectrum　transform　algorithm　can　identify　and　extract　fault　features,　which　provides　a　theoretical　basis　for　the　fast　detection　and　location　of　short-circuit　faults　in　microgrid　in　the　future.　©　2019　IEEE.