The　uncertainty　of　wind　power　brings　security　and　economic　issues　in　active　distribution　networks.　As　an　effective　solution　to　these　problems,　the　active/reactive　optimization　aims　at　minimizing　the　total　operation　cost　and　network　loss　by　controlling　multiple　resources.　This　paper　proposes　a　two-stage　data-adaptive　robust　optimization　model　for　the　dynamic　active-reactive　power　flow.　The　dynamic　correlation　between　the　uncertainties　of　active　and　reactive　power　is　analyzed　at　two　kinds　of　power　factor　modes　of　wind　turbines　and　then　is　described　by　a　linear　affine　decision　rule.　To　characterize　these　uncertainties,　a　data-adaptive　multi-band　uncertainty　set　is　constructed　based　on　historical　data,　which　allows　the　number　of　bands　and　the　corresponding　weight　coefficients　to　be　adjusted　under　different　confidence　levels.　Subsequently,　the　conic　relaxation　and　linearization　techniques　are　performed　to　reformulate　the　robust　optimization　model　as　a　mixed-integer　second-order　cone　programming　problem.　A　two-level　solution　framework,　consisting　of　a　column-and-constraint　generation　algorithm　and　an　outer　approximation　algorithm,　is　employed　to　solve　this　problem.　Simulation　results　on　a　modified　IEEE　33-bus　system　indicate　that　the　variable　power　factor　mode　is　more　suitable　for　describing　the　dynamic　correlation　than　the　constant　power　factor　mode.　Besides,　the　proposed　model　and　method　also　show　better　practicality　and　solution　efficiency　than　the　existing　ones.　©　2022　Elsevier　Ltd