It＇s　well-known　that　dead　reckoning　algorithm　(DR　algorithm)　used　in　distributed　interactive　simulation　(DIS)　can　effectively　reduce　the　amount　of　communication　and　its　synchronization　accuracy　is　acceptable　commonly.　But　in　the　scenario-based　interactive　simulation　applications,　when　DR　algorithm　is　used　alone　the　state　of　nodes　are　still　likely　to　be　inconsistent　because　of　network　delay　and　user　interaction.　And　that　would　lead　to　some　conflicts　in　the　state　of　key　events　and　scenes.　A　new　practical　method　is　proposed　to　solve　this　problem　by　using　the　adaptive　state　update　strategy　in　sensitive　region.　The　state　update　frequency　can　be　adaptively　adjusted　according　to　the　congestion　state　of　the　network.　The　distributed　interactive　simulation　for　explosive　disposal　training　is　conducted　to　test　our　method.　Experimental　results　show　that　it　can　effectively　maintain　the　real-time　state　synchronization.　©　2011　IEEE.