Cognitive radio is a promising technology for alleviating the spectrum shortage problem which is caused by current fixed spectrum allocation policy. With a smart antenna configured in a secondary user, this paper proposes a newspectrum sensing algorithm exploiting the correlation of channel gains between antenna elements. The proposed method detects the primary user signal with the test-statistics based on antenna correlation. Compared to available antenna correlation based spectrum sensing algorithm, the proposed algorithm can determine easily the decision threshold level for achieving required false alarm probability under given number of samples, and achieves a better performance. Finally, the proposed algorithm is verified by numerical simulation. determine easily the decision threshold level for achieving required false alarm probability under given number of samples, and achieves a better performance. Finally, the proposed algorithm is verified by numerical simulation.