Abstract: When cognitive radio system is under the deep shadowing and fading environment, cooperative spectrum sensing can improve the performance of spectrum sensing by fully using multiple cognitive radio user’s spatial diversity gain. The cooperative spectrum sensing can reduce the demand of single cognitive user’s sensitivity and decrease the system misjudgment resulted from the single detection uncertainty. Some studies show that cooperative spectrum sensing based on the D-S evidence theory has better SNR robustness and the detection performance than traditional sensing method, e.g. OR rule, AND rule and Optimal fusion rule. However, the conflicts information limits the performance of D-S method. In this paper, a novel scheme of cooperative spectrum sensing based on discounted DSmT was proposed to treat the conflicts information, which discounts the basic probability assignment of the terminals according to the credible degree of data from each local spectrum sensing terminal. Our method can add the fusing contribution of high reliability data, and reduce the interference of the low reliability data to the fusion results. Simulation results show that our method has the best detection performance than the traditional combination rule for cooperative spectrum sensing.