Due to advances in communications technology in recent years, prompting more extensive application of wireless sensor networks. Such sensors are limited in battery energy supply and produce the energy hole problem, this even causes paralysis of part of the system. Therefore, the energy consumption becomes the subject of this thesis discusses. In the cluster architecture, burden of the cluster head is bound to become the energy consumption of the maximum point, so our method is focused on reducing energy consumption of the cluster head. To this end, we propose a new scheduling mechanism based on cluster architecture. In this mechanism, we use the "polling" method to make the cluster head have an absolutely effective data receiving. In addition, we also introduced the "sleeping" mechanism to ensure that the cluster head can achieve power saving under the premise of the most effective data receiving.