The demand for wireless communication has grown remarkably in the last year, consequently raising the problem of spectrum scarcity. In this context, cognitive radio is an emerging technology that aims to overcome that scarcity which is one of the most challenging problems in modern wireless communication. Among its fundamental function, the most important is the spectrum sensing which require precise accuracy and low complexity. Thus, various signal detection methods have been proposed for multiresolution spectrum sensing (MRSS). None of these techniques have been used in a wavelet based cyclostationary feature detector. To achieve that we suggest a cyclostationary feature based MRSS in the context of IEEE 802.22 Wireless Regional Area Network (WRAN) for cognitive radio to classify and identify the primary signal either Digital Video Broadcasting- Terrestrial (DVB-T) or wireless microphone signal. This knowledge of identifying primary signals can help cognitive radio to use fraction of TV band when only a wireless microphone signal is present in the channel. The performance of the proposed scheme is evaluated by probability of correct classification. The result indicates that better performance can be achieved by the proposed scheme especially in a low SNR environment. © 2010 IEEE.