In this paper, we propose a decoupled substation price model with separate charges for base energy production, up/down reserve usages and reserve usage variations. Using the decoupled price scheme, the flexible loads in the distribution system are scheduled to closely follow variations in renewable generation in order to mitigate demand fluctuations in substation. This reduces the need for reserve unit power production, thereby allowing transmission systems to operate at lower cost and higher efficiency. We formulate an optimization problem to determine the optimal scheduling of flexible loads subject to power balance, power flow and demand response constraints. The optimization problem is solvable by computationally efficient linear programming methods. Realtime and day-ahead models are considered. Numerical examples on a modified 13-node test feeder demonstrate the effectiveness of the proposed method.