New wireless communication technologies increase the demand for spectrum sensing. However due to the limitations of analogue-to-digital converters, sampling a wide spectrum still remains as a major problem. In this paper, a method for estimating the power spectrum density (PSD) from multi-coset (MC) samples, which is a special sub-Nyquist sampling method, is explained. In order to see the feasibility of the method, wide-sense spectrum measurements are taken from a vector signal analyzer and each measurement is MC sampled. Then the PSD estimation is compared with the original PSD of the signal. False alarm rates for different variations are analyzed.