Existing Collaborative Spectrum Sensing (CSS) algorithms in the presence of Spectrum Sensing Data Falsifi cation (SSDF) attacks have been investigated in the small
scale attacks, where malicious users are assumed to be in a minority and have limited effects on fi nal decision. But, in massive attacks, where there are a large number of malicious users, the fi nal decision is unreliable and existing methods have low effectiveness. In contrary, we propose a new Weighted Likelihood Ratio Test (W LRT) that collaborative weight is calculated by comparing the sensing history of each user with the reliable anchor nodes' global decision. The obtained weights are applied in LRT to improve the CSS performance. Simulation results verify the eff ectiveness of the proposed method.