Abstract: In a cognitive radio network, the channel condition state and user state which are both strong time-varying affect the system performance. In order to improve the system throughput and achieve the maximum communication rate of secondary users (SUs), how to match the channel state and user state with adaptive modulation and coding (AMC) scheme is an urgent problem to be solved. This paper proposes a matching-check allocation protocol (MAP) which combines adaptive packet error rate and communication rate to maximize the total throughput of the system while matching the channel state and user state with AMC scheme. In order to evaluate the performance of spectrum allocation protocol under complex network, such as multi-user multi-channel joint objective optimization, this paper using discrete-time queueing model analyses the behavior of SUs to construct a universal and applicable protocol evaluation framework, which obtains performance indexes, such as the average queue length, throughput, delay and rejection. In this framework, MAP protocol is compared with maximum-rate allocation protocol (MRP) and fair-random allocation protocol (FAP). Numerical results show that MAP is superior to MRP and FAP, and the protocol evaluation framework can evaluate different spectrum allocation protocols, which proves the effectiveness of MAP protocol and protocol evaluation framework.