Limited　resources　are　a　prevalent　and　challenging　problem　in　the　field　of　emergency　management.　Emergency　scheduling　is　an　effective　way　to　make　full　use　of　resources.　However,　designing　an　effective　emergency　plan　to　minimize　rescue　time　is　a　major　challenge.　This　study　focuses　on　large-scale　emergency　scheduling　for　fighting　forest　fires　with　multiple　rescue　centers　(depots)　and　limited　fire-fighting　resources,　which　aims　to　determine　the　optimal　rescue　route　of　fire-fighting　teams　at　multiple　rescue　centers　to　minimize　the　total　completion　time　of　all　fire-fighting　tasks.　For　this　problem,　we　first　assign　rescue　priorities　to　different　fire　points　according　to　the　speed　of　the　fire　spread.　Then,　we　formulate　it　into　a　mixed-integer　linear　programming　(MILP)　model　and　analyze　its　NP-hard　complexity.　To　deal　with　large-scale　problems,　a　new　fast　and　effective　artificial　bee　colony　algorithm　and　variable　neighborhood　search　combined　algorithm　is　proposed.　Extensive　experimental　results　for　large-scale　randomly　generated　instances　confirm　the　favorable　performance　of　the　proposed　algorithm　by　comparing　it　with　MILP　solver　CPLEX,　genetic　algorithms,　and　particle　swarm　optimization　algorithms.　We　also　derive　some　management　insights　to　support　emergency　management　decision-making.