Spectrum sensing is a key task for cognitive radio. Our motivation is to increase the probability of detection for spectrum sensing in cognitive radio. In this paper, we proposed a new semi blind method which is based on minimum Eigenvalue of a covariance matrix. The ratio of the minimum eigenvalue to noise power is used as the test statistic. The method does not need channel and signal information as prior knowledge. Eigenvalue based algorithm perform better than energy detection for correlated signal. Our proposed method is better than the maximum eigenvalue and energy detection for correlated signal. We perform Simulation which is based on digital TV signal. In all tests, our method performs better than maximum eigenvalue detection and energy detection. DOI: http://dx.doi.org/10.11591/ijece.v4i4.6226