Abstract: A suitable energy management mechanism for a wireless network with intermittent connectivity was proposed to deal with unbalanced load and limited energy resources of nodes. According to the historical information, the active degree, residual energy and data forwarding rate of nodes were estimated in a distributed way. In addition, the node utility was effectively estimated by fully considering network features. Furthermore, the proposed mechanism employs the serviceability differences of nodes and the Pareto optimal theory to choose the best next-hop relay node adaptively. The data forwarding operation was executed. Thus the network performance degradation caused by selfish nodes was prevented effectively. The simulation results show that the proposed mechanism can not only balance the load on nodes and solve the problem of network "hotspots", but also prolong network lifetime and improve its delivery rate and delay performance greatly compared to other energy management mechanisms.