The　extensive　installation　of　gas-fired　units　and　the　integration　of　large-scale　wind　power　connected　to　power　grid　have　strengthened　the　interdependency　between　power　system　and　natural　gas　network.　Consequently,　wind　power　uncertainty　accompanying　with　the　coupling　interaction　by　gas-fired　units　has　brought　new　challenges　to　the　safe　and　economic　operation　of　electricity-gas　coupled　system.　A　decomposition-coordination　framework　is　developed　to　study　the　cooperative　optimization　operation　of　the　integrated　electricity-gas　coupled　system.　In　this　framework,　a　data-driven　distributionally　robust　optimization　(DDRO)　model　is　proposed　to　solve　the　power　system　scheduling　problem　with　wind　power　uncertainty.　Aiming　to　minimize　the　expectation　of　the　operation　cost　under　the　worst-case　distribution,　DDRO　combines　the　advantages　of　the　traditional　robust　optimization　(RO)　and　stochastic　optimization.　Case　studies　are　implemented　on　two　electricity-gas　coupled　systems　of　different　scales　to　verify　the　effectiveness　of　the　proposed　decomposition-coordination　framework　with　DDRO.　Specifically,　compared　with　the　distributionally　robust　optimization　(DRO)　model　based　on　moment　information,　the　solution　obtained　by　DDRO　can　save　1765.01　$　for　the　6-bus　power　system　and　2331.18　$　for　the　IEEE　24-bus　power　system,　respectively.　It　is　demonstrated　that　DDRO　can　achieve　less　conservative　and　more　economic　scheduling　solutions　compared　to　DRO.　©　2019　Elsevier　Ltd