The　uncertainty　of　source　and　load　along　with　the　wind　power　correlation　can　make　the　scheduling　results　of　regional　integrated　energy　system　less　credible.　To　solve　this　problem，taking　the　minimum　comprehensive　operation　cost　as　the　objective，a　scheduling　model　of　regional　integrated　energy　system　with　multi-time　scale　is　proposed.　In　the　day-ahead　stage，the　two-stage　robust　optimization　model　considering　wind　power　correlation　is　proposed，and　the　column-and-constraint　generation　method　is　used　for　the　iterative　solution.　In　the　intra-day　scheduling　stage，the　different　response　rates　of　cold，heat　and　power　are　considered，and　the　hierarchical　rolling　optimization　model　of　cold，heat　and　power　based　on　model　predictive　control　is　proposed　to　further　eliminate　the　fluctuation　of　source　and　load　power.　The　simulative　results　show　that　the　robust　optimization　method　considering　the　wind　power　correlation　reduces　conservatism　and　improves　economy.　The　economic　and　stable　operation　of　regional　integrated　energy　system　is　realized　by　using　model　predictive　control　in　the　hierarchical　optimization　of　cold，heat　and　power.　©　2023　Electric　Power　Automation　Equipment　Press.　All　rights　reserved.