Abstract: Dynamic transmission power control to the transmitters of the cognitive radio links is accomplished by a new iterative algorithm. The power and cost of spectrum utility lost are constrained with each other in the iteration; the signaltointerference plus noise ratios to the signal received are changed as the iteration progressing, so as to obtain the optimal transmit power distribution to use the spectrum rationally. In order to reduce the interference among the cognitive radio links, an access screening mechanism is proposed. The interference probability distribution is used to describe the uncertainty of the interference of a cognitive radio link. The information entropy corresponding to the interference probability is calculated to determine the adaptive interference threshold required in the access screening mechanism. Simulation shows that the total interference in the system reduces from 3 mW to 0.5 mW using the proposed scheme when there are 40 cognitive radio links.