Costly　and　time-consuming　maintenance　of　the　hydrogen　compressors　due　to　their　frequent　breakdown　severely　hinders　the　deployment　and　promotion　of　hydrogen　refueling　station　(HRS),　and　effective　condition　monitoring　and　fault　diagnosis　is　the　key　to　reduce　the　unscheduled　downtime　of　the　compressor.　This　paper　proposes　a　nondestructive　method　for　fault　diagnosis　of　diaphragm　compressors　for　HRSs　based　on　the　acoustic　emission　(AE)　signal.　The　AE　signals　in　the　time　domain　are　segmented　into　angle　domain　signals　correspond　to　a　working　cycle.　The　feature　events　of　the　moving　components　are　determined　through　the　measured　AE　signal　in　both　angle-domain　and　angle　frequency　domain　based　on　short-term　Fourier　transform　(STFT).　Those　feature　events　signals　are　innovatively　applied　to　identify　the　typical　abnormal　conditions　of　excessively　high　oil　pressure,　slightly　inadequate　oil　pressure　and　seriously　inadequate　oil　pressure,　replacing　the　traditional　and　destructive　pressure　measuring　method.　The　results　show　that　this　method　can　be　used　to　effectively　diagnose　abnormal　working　conditions　and　indicate　that　this　method　can　be　utilized　as　a　powerful　tool　in　the　non-destructive　condition　monitoring　and　fault　diagnosis　of　the　diaphragm　compressors.　(C)　2019　Hydrogen　Energy　Publications　LLC.　Published　by　Elsevier　Ltd.　All　rights　reserved.