This　paper　employs　Data　Warehouse　theory,　theory　of　Decision　Support　System,　and　the　ASP.net　technology　to　implement　the　Power　Marketing　DSS　(Decision-making　Support　System)　and　put　forward　the　realization　of　entire　system　architecture.　In　the　data　layer,　various　data　from　Applications　which　are　running　for　the　Power　Business　are　integrated　to　establish　the　data　warehouse,　and　then　we　build　the　multi-dimensional　data　cube　to　provide　online　analytical　processing.　In　application　layer,　the　Webservice　technology　as　well　as　its　Registration　Model　and　query　technology　based　on　UDDI　registry　will　be　introduced　to　Power　consumption　forecasting　module.　The　forecasting　models　include　Quadric　Exponential　Smoothing　prediction　model,　Curve　Fitting,　Multiple　Linear　Regression　Model　and　the　Gray　Model,　each　model　will　be　packaged　into　a　Web　service,　so　the　models　can　be　called　through　the　network　for　Distributed　Collaborative　Forecasting.　Through　these　technologies,　a　complete　and　well-established　Power　Marketing　DSS　was　created　with　following　functions:　on-line　analysis,　report　service,　distributed　collaborative　power　forecasting　models　work　in　parallel,　etc.