Abstract: Green Communications is an inevitable trend of future development of wireless communication. In order to enhance the stability and efficiency of spectrum sensing algorithm and promote low-power Green Communications, the energy efficiency of eigenvalue-based MME (Maximum-Minimum Eigenvalue) algorithm under noise uncertainty is analyzed in this paper and the optimal sensing time through the golden algorithm is found. Theoretical analysis and computer simulation results show that eigenvaluebased MME detection algorithm is obviously better than energy detection(ED) with the noise uncertainty and has a more stable sensing performance at low signal to noise ratio. Meanwhile,the results also indicate that noise uncertainty almost has no effect on MME.