This　paper　proposes　a　new　method　for　dynamic　airspace　configuration　based　on　a　weighted　graph　model.　The　method　begins　with　the　construction　of　an　undirected　graph　for　the　given　airspace,　where　the　vertices　represent　those　key　points　such　as　airports,　waypoints,　and　the　edges　represent　those　air　routes.　Those　vertices　are　used　as　the　sites　of　Voronoi　diagram,　which　divides　the　airspace　into　units　called　as　cells.　Then,　aircraft　counts　of　both　each　cell　and　of　each　air-route　are　computed.　Thus,　by　assigning　both　the　vertices　and　the　edges　with　those　aircraft　counts,　a　weighted　graph　model　comes　into　being.　Accordingly　the　airspace　configuration　problem　is　described　as　a　weighted　graph　partitioning　problem.　Then,　the　problem　is　solved　by　a　graph　partitioning　algorithm,　which　is　a　mixture　of　general　weighted　graph　cuts　algorithm,　an　optimal　dynamic　load　balancing　algorithm　and　a　heuristic　algorithm.　After　the　cuts　algorithm　partitions　the　model　into　sub-graphs,　the　load　balancing　algorithm　together　with　the　heuristic　algorithm　transfers　aircraft　counts　to　balance　workload　among　sub-graphs.　Lastly,　airspace　configuration　is　completed　by　determining　the　sector　boundaries.　The　simulation　result　shows　that　the　designed　sectors　satisfy　not　only　workload　balancing　condition,　but　also　the　constraints　such　as　convexity,　connectivity,　as　well　as　minimum　distance　constraint.　(C)　2014　Production　and　hosting　by　Elsevier　Ltd.　on　behalf　of　CSAA　&　BUAA.