Abstract: Two energy-efficient strategies are proposed to reduce the energy consumed in the spectrum sensing procedure for wireless cognitive sensor networks. One involves using confidence voting algorithm to reduce the detection number, the other is based on Partially Observable Markov Decision Process (POMDP). Simulation results show that both algorithms can reduce the energy consumed in the detection efficiently, and the combination of the two algorithms can save more than 30% energy compared with the general algorithm with system detection reliability requirement guaranteed.