The　green　supply　chain　realizes　the　unification　of　economic　and　environmental　benefits　through　green　manufacturing,　green　circulation　and　reverse　logistics,　and　represents　an　important　way　to　reduce　emissions.　The　characteristics　of　the　green　supply　chain　network,　however,　can　encourage　some　companies　to　exhibit　＂free-riding＂　behavior,　participating　without　being　willing　to　reduce　emissions　directly　themselves.　This　paper　applies　an　evolutionary　game　model　to　a　two-level　green　supply　chain　composed　of　green　suppliers　and　green　manufacturers　in　order　to　analyze　a　variety　of　internal　and　external　factors　that　affect　the　behavior　of　both　parties　in　the　game,　and　thence　numerically　simulate　the　evolution　and　stability　trend　of　coordinated　reductions　in　emissions.　The　results　identify　many　cases　of　system　evolution　but　the　only　stable　evolution　strategy　is　when　both　(1)　the　sum　of　collaborative　emission　reduction　benefits　and　government　subsidies　is　greater　than　the　sum　of　collaborative　emission　reduction　input　costs　and　＂free　rider＂　benefits,　and　(2)　the　increased　rate　of　unilateral　emission　reduction　benefits　is　greater　than　the　ratio　of　costs　to　original　benefits.　Income　arising　from　collaborative　reductions　in　emissions,　enterprise　original　income,　income　from　unilateral　reductions　in　emissions　increase　ratio,　government　subsidy　coefficient　and　achievement　reward　base　can　directly　affect　the　system　evolution　path.　The　larger　these　value,　the　greater　the　probability　of　green　suppliers　and　green　manufacturers　choosing　collaborative　reductions　in　emissions,　and　the　faster　the　system　convergence　speed.　Only　the　imposition　of　regulatory　punishments　above　the　threshold　will　affect　the　＂free　riding＇　behavior　of　enterprises　and　drive　the　upstream　and　downstream　enterprises　of　the　green　supply　chain　to　reduce　emissions　faster.