Abstract : Due to the rapid proliferation of wireless technologies, spectrum management has become a serious concern in 5G digital world. According to the studies of Federal Communication Council, there are many areas in spectrum which are unused. Cognitive Radio uses innovative spectrum sensing and sharing techniques to allocate unused channels with secondary users. For efficient sharing of resource, sensing of an idle spectrum is necessary. In this paper, the spectrum sensing is improved using a statistical method, where the idle state of additive white Gaussian channel is estimated using the fourth order moment of the windowed signals defined as kurtosis. The proposed work improves probability of detection (Pd) as kurtosis doesn't depend on the energy of the signal. When received signals undergo signal processing like Fast Fourier transform, the spectrum increases leading to false detection. But in this work, probability of miss detection (Pmd) is reduced as the detection is based on windowed FFT based Kurtosis. © 2018 IEEE.