Wireless communication is an emerging and evergreen field of research and will continue to be so in the future. Increased demands in spectrum and users eventually result in spectral congestion. Cognitive Radio systems provide an intelligent and an attractive solution to this problem. Spectrum Sensing is the main key for the functioning of Cognitive Radio and helps in efficient spectrum utilization. Wavelets provide a new dimension to tackle spectrum sensing when compared to the conventional methods. In real time applications, loss of signal is prominent in low SNR conditions and hence a system capable of working under such conditions will prove to be an asset. In this paper, a wavelet based approach to spectrum sensing at low SNRs is dealt with. A new algorithm for spectrum sensing has been proposed here which emphasizes more on the functioning of systems in low SNR conditions.