To analyze the effect of demand response (DR) in the regional power grid, a day-ahead optimal dispatching model of the regional power system considering the DR is established by using the elasticity matrix of electricity price, and a real-time rolling dispatching optimization model considering the renewable energy power prediction errors in different time scales is established. For the privacy of information in different regions, the alternate direction multiplier method (ADMM) is used for a distributed solution to obtain an optimal scheduling scheme. Taking a regional power grid in China as a case, we testify that the DR plays a crucial part in peak-load shifting in the regional power grid. Under the reasonable proportion of curtailable load (CL) and shiftable load (SL), the dispatching optimization model proposed can promote the total operation cost reduction and renewable energy consumption.