Cognitive radio is a promising technology to increase spectrum utilization. In this technology, cognitive radios (secondary users) share frequency spectrum with licensed users (primary users). To enable spectrum sharing, cognitive radio supports two functionalities: spectrum sensing and primary users classification. In this paper we evaluate the performance of cyclostationarity-based classification approach in multipath fading channel and noise uncertainty. This approach was proposed in context of WRAN 802.22; the first cognitive radio standard. The results show that the proposed approach performs well to recognize TV-PAL and wireless microphone as the primary users. © 2010 IEEE.