Sleep Scheduling is one of the most important challenges for energy conservation in Wireless Sensor Networks (WSNs). Connected Dominating Set (CDS) based virtual backbone is often employed in WSNs. Compared with the ordinary CDS, the R-hop Connected Dominating Set (R-CDS) with smaller size is more suitable for build a virtual backbone for energy conservation. In this paper, a R-hop connected dominating sets based Partition Algorithm (RPA) is proposed to construct the R-CDP with the knowledge of local connectivity information only. RPA establishes multiple overlapped virtual backbones for WSNs such that every virtual backbone works in sleep schedule to achieve energy consumption evenly among all sensor nodes for prolonging the lifetime of the networks to maximization. The simulation results show that compared with the other schemes, our strategy can improve the performance of the networks in longer lifetime.