The　river-blocking　landslide　disasters　are　widely　distributed　in　the　mountainous　areas　of　southwest　China,　characterized　by　high-elevation　long-runout　movements　with　significant　destructive　power.　The　identification　and　monitoring　of　high-elevation　long-runout　landslides,　as　well　as　the　prediction　of　unstable　landslide　movements,　hold　great　significance　for　regional　disaster　mitigation　and　prevention.　In　this　study,　we　used　interferometric　synthetic　aperture　radar　to　identify　and　monitor　potential　landslides　in　the　Gongjue　segment　of　the　Jinsha　River　Basin　in　China.　The　Sela　landslide　was　selected　for　rainfall　infiltration　simulation,　predict　the　entire　process　of　landslide　instability,　and　explore　its　failure　characteristics　from　a　dynamic　perspective.　The　monitoring　results　revealed　the　presence　of　four　typical　landslides　in　Gongjue　County,　situated　within　the　Jinsha　River　area,　with　slope　deformation　rates　exceeding　17　cm/yr.　The　maximum　observed　slope　deformation　rate　reaches　46　cm/yr.　Decomposition　results　of　the　time-series　deformation　characteristics　of　the　landslide　feature　measurement　points　show　that　rainfall　was　the　primary　factor　influencing　the　periodic　deformations　of　landslides.　Simulation　results　indicated　that　rainfall　promotes　landslide　deformation.　Under　the　influence　of　rainfall,　the　movement　speed　of　the　landslide　increases　rapidly,　and　the　front　sliding　body　slides　first,　followed　by　the　instability　of　the　upper　sliding　body.　This　is　consistent　with　the　damage　mode　caused　by　a　traction-type　landslide.　Subsequently,　the　particle　movement　speed　maintains　a　short　period　of　rapid　sliding,　gradually　decreasing　and　eventually　reaching　a　stable　state.　The　results　of　this　study　provide　an　important　reference　for　carrying　out　remote　monitoring　and　risk　prediction　for　high-elevation　long-runout　landslides.