The　increasing　expansion　of　wind　and　gas　turbine　installation　has　intensified　the　interdependency　between　power　system　and　natural　gas　network,　which　poses　great　challenges　to　the　coordination　scheduling　of　the　multi-energy　coupled　system　(MECS).　The　data-driven　robust　optimization　(DDRO)　model　is　proposed　for　the　energy　coupled　system.　In　this　model,　the　temporal-spatial　correlation　of　wind　power　can　be　considered　based　on　the　minimum　volume　enclosing　convex　hull　uncertainty　set,　and　the　confidence　set　about　the　probability　distribution　for　wind　power　scenarios　with　the　form　of　the　norm-1　and　norm-inf　constraints　is　constructed　to　handle　wind　power　uncertainty.　Moreover,　to　describe　natural　gas　transient　characteristics,　the　hydrodynamic　model　for　gas　flow　represented　as　a　series　of　partial　differential　equations　is　transformed　by　the　Wendroff　difference　scheme　and　linearization　technique.　And　then　the　master-subproblem　framework　and　tri-level　duality-free　decomposition　method　is　developed　to　solve　the　above　model.　Finally,　the　proposed　model　and　solving　method　are　carried　out　on　two　test　systems　with　different　scale,　and　the　robust　optimization　models　and　distributionally　optimization　models　in　the　existing　literatures　are　implemented　for　comparison.　Simulation　results　demonstrate　the　effectiveness　and　superiority　of　the　proposed　model　for　solving　the　coordination　scheduling　problem　of　MECS.　(C)　2020　Elsevier　Ltd.　All　rights　reserved.