The　forest　fires,　which　occur　around　the　world　each　year,　have　had　a　great　influence　on　air　quality.　It　is　difficult　to　research　the　pollutant　contribution　of　forest　fire　smoke　long-distance　transport　to　the　ambient　environment.　This　study　attempted　to　use　numerical　models　to　investigate　the　possible　air　quality　effect　of　Great　Xing＇　anling　forest　fire　smoke　on　other　regions　from　28　July　to　18　August　in　2002.　The　meteorological　model　MM5　(Fifth-Generation　NCAR　/　Penn　State　Mesoscale　Model)　was　used　to　analyze　the　synoptic　pattern　during　the　fire　episode　and　supply　the　track　analysis　model　with　meteorological　field.　MM5　was　configured　with　domain　covering　China　of　105-135　degrees　E,　35-60　degrees　N,　with　horizontal　grid　spacing　of　30km.　Twenty　four　full　sigma　levels　extend　from　surface　to　the　model　top　of　200　hpa.　It　was　found　that　there　was　a　long-term　stationary　high　pressure　system　controlling　most　regions　of　North　China　during　5-14　Aug,　which　was　conductive　to　the　storage　of　air　pollutants　and　made　it　keep　a　relatively　high　concentration　during　the　slowly　long-distance　transport.　Besides,　the　Hybrid　Single-Particle　Langrangian　Integrated　Trajectory　model　version　4　(HYSPLIT_4)　was　used　in　forwards　mode　to　make　a　research　on　the　possible　affected　regions　of　the　forest　fire　pollutants.　The　horizontal　track　analysis　results　indicated　that　regions　of　north　Inner　Mongolia　were　all　affected　by　the　fire　smoke.　The　comparison　between　vertical　track　analysis　and　ground　observations　of　PM10　concentration　in　Beijing　suggested　that　the　heavily　polluted　PM10　released　from　the　forest　fire　may　had　made　certain　contributions　to　the　high　PM10　level　in　Beijing　during　this　period.　The　methodology　of　forest　fire　smoke　long-distance　transport　track　analysis　based　on　MM5　and　HYSPLIT_4　makes　three-dimensional　forest　fire　transport　analysis　come　true,　and　it　would　be　helpful　to　analyze　the　transport　path　of　smoke　in　sudden　fire　accidents.