In　order　to　study　the　temporal　and　spatial　changes　and　relations　between　the　conductivity　of　soil　slope　and　matrix　suction　under　the　action　of　rainfall,　a　physical　model　test　of　artificial　rainfall　landslide　was　conducted,　in　which　the　resistivity　was　used　as　the　monitoring　quantity,　combined　with　the　conventional　water　content　and　matrix　suction,　and　Manila　grass　was　planted　on　the　slope　model　to　explore　the　infiltration　characteristics　of　grass　slope　under　three　dynamic　rainfall　modes:　uniform　rainfall,　front　small　back　peak,　and　front　peak　back　small.　The　results　show　that　under　the　dynamic　rainfall　mode,　the　change　of　rainfall　intensity　can　affect　the　resistivity　and　water　content　of　soil,　but　there　is　a　certain　lag.　Although　the　response　time　of　soil　matric　suction　is　different　under　different　rainfall　patterns,　the　variation　trend　of　soil　matric　suction　is　similar.　Before　rainfall,　the　matrix　suction　and　resistivity　of　slope　soil　are　higher,　and　gradually　decrease　with　the　increase　of　depth.　After　the　rainfall,　due　to　water　infiltration,　the　matric　suction　and　resistivity　of　surface　soil　decreased　greatly,　and　the　distributions　on　the　slope　have　showed　a　trend　of　increasing　gradually　from　shallow　layer　to　deep　layer.　A　prediction　model　of　residual　soil　matric　suction　based　on　resistivity　was　obtained,　combined　with　Keller＇　s　improved　Archie　model　and　VG　model,　which　was　verified　by　experimental　data,and　the　results　were　reasonable.　This　model　adds　a　new　way　for　indirect　measurement　of　residual　soil　matrie　suction.　The　research　results　are　helpful　to　explore　the　seepage　characteristics　of　unsaturated　residual　soil　slope　under　different　rainfall　modes,　and　reveal　the　evolution　law　of　the　electrical　conductivity　and　spatial　distribution　characteristics　of　matrie　suction　of　residual　soil　slope　under　rainfall,　which　is　of　theoretical　and　practical　significance　for　stability　analysis,　monitoring　and　early　warning　of　residual　soil　slope　under　typhoon　and　rainstorm.　©　2024　Institute　of　Engineering　Mechanics　(IEM).　All　rights　reserved.