This　paper　studies　an　online　scheduling　problem　in　the　smart　grid,　which　is　arised　in　demand　response　management　under　the　scenario　with　real-time　communication　between　the　grid　operator　and　consumers.　Consumers　send　the　power　requests　online　over-list.　The　request　is　released　with　a　limited　set　of　timeslots.　Only　one　of　the　timeslots　in　the　set　can　this　request　be　served　by　the　operator.　In　a　timeslot,　the　electricity　cost　consumed　to　serve　the　requests　is　a　quadratic　function　of　the　load　in　it.　Our　aim　is　to　find　a　best　possible　online　schedule　which　generates　the　minimal　total　electricity　cost.　In　this　paper,　we　propose　a　greedy　algorithm　of　this　problem　which　is　2-competitive.　Besides,　we　prove　our　algorithm　is　optimal.