Off-grid　microgrids　lack　of　backup　of　the　main　network,　which　poses　a　higher　challenge　to　the　effective　energy　supply,　how　to　improve　the　matching　degree　of　supply　and　demand　has　become　a　popular　direction　of　current　research.　Therefore,　firstly,　on　the　basis　of　considering　the　bilateral　uncertainty　of　supply　and　demand,　based　on　the　demand　response　model,　a　chance-constrained　bi-level　optimal　dispatching　model　is　constructed　to　promote　the　linkage　of　＂demand-supply＂;　Then,　considering　that　the　demand　side　participation　in　the　demand　response　on　the　energy　supply　side　can　bring　incremental　benefits,　an　benefit　allocation　strategy　for　both　sides　of　supply　and　demand　is　constructed;　Finally,　an　off　grid　microgrid　in　Northwest　China　is　selected　as　the　research　background,　using　scenario　generation　and　opportunity　constraints　to　deal　with　uncertainty,　and　using　an　improved　genetic　algorithm　to　solve　the　bi-level　optimal　dispatching　model.　The　results　show　that:　The　proposed　two-level　un-certainty　processing　model　can　effectively　reduce　the　energy　supply　deviation;　The　optimal　dispatching　model　proposed　can　realize　＂demand-supply＂　interaction　and　reduce　the　load　shedding　rate;　The　installed　capacity　of　units　can　be　increased　to　further　reduce　the　load　shedding,　but　it　will　cause　higher　curtailment　of　clean　energy,　consideration　can　be　given　to　implementing　demand　response　or　increasing　energy　conversion　units　to　enhance　efficiency　of　clean　energy　utilization;　The　improved　Shapley　value　method　can　improve　the　willingness　of　the　demand　side　to　actively　participate　in　the　demand　response　of　the　supply　side,　and　improve　the　satisfaction　of　the　cooperative　alliance.