Abstract: For multi-channel cognitive radio network (CRN), the inconsistent transmission conditions of common control channel (CCC) and data channels deteriorate the performance of cognitive MAC protocols. Focusing on this problem, a novel channel aware multi-channel cognitive MAC protocol (CAM-MAC) was proposed, which consists of two-level four-way handshaking mechanism. The first level two-way handshaking was optimized to reduce the mean successful reservation time on CCC. The other level two-way handshaking was implemented by a cross-layer approach to adopt the adaptive transmission based on instantaneous SNR on each Nakagami fading data channel to improve the data rate. These mechanisms improve the performance of CAM-MAC. Simulation results validate the throughput of CAM-MAC outper-forms that of the existing one, about 50% improvement in the throughput under certain conditions.