In　order　to　purify　Rayleigh-wave　signals　fro　m　seis　mic　prospecting　records,　apply　Gaussian　neighboring　threshold　filtering　to　Gaussian　Rayleigh-wave　signal　purificat　ion　study　according　to　area　selection　issue　in　ST　time-frequency　filtering.　Based　on　different　time-frequency　distribution　types,　three　different　Gaussian　neighboring　threshold　filtering　models　are　defined.　Local　threshold　processing　is　conducted　on　signals　in　S　matrix　Gaussian　neighboring　time-frequency　data　based　on　different　models　of　Gaussian　neighborhoods.　Preserve　useful　information,　and　obtain　filtering　signals　after　general　s-transform.　Results　show　that　GST　neighboring　threshold　filtering　method　has　high　filtering　precision　and　the　filtering　model　can　effectively　reduce　filtering　mean　square　error.