In view of cheating behavior of the secondary user (SU) affecting the performance of non-cheating SU in the process of cognitive radio channel allocation, a channel allocation protocol named the improved top trading cycles (ITTC) is proposed according to the top trading cycle (TTC) matching theory. The ITTC allocates channels according to the preference ranking of SUs, and the behavior of changing the ranking of the true preference list is in consideration of cheating. The cheating SUs can only obtain worse allocation performance, while non-cheating SUs can obtain the same or even better allocation results than before, effectively solving the unfairness caused by cheating. For quantitatively analyzing the allocate performance indicators, a performance evaluation method of SU (PEM-SU) is designed. PEM-SU configures the PU network, as well as the arrival, distribution, and transmission of SU, and unifies the results into a form that can be quantified for performance. The experimental results compared with the secondary proposed deferred acceptance (SPDA) protocol and TTC protocol show that the ITTC is prior to SPDA and TTC in the average throughput, average latency, average queue length and average rejection rate. And also can effectively prevent strategic manipulation and protect the interests of non-cheating SUs.