Cognitive radio is able to effectively solve problems like spectrum resource scarcity and low spectrum utilization rate in wireless communication system. The detection performance of multi-bit decision algorithm is much better than that of conventional 1-bit decision algorithm. Aiming at attacks from malicious users in multi-bit decision model when cognitive users are mobile, this paper proposes a multi-bit decision cooperative spectrum sensing algorithm in mobile scenarios based on trust valuations. It adopts two kinds of reliabilities: location reliability and user reliability. The former is used to describe average fading characteristics and average path loss characteristics of wireless channels in different areas and the latter is used to describe reliability of each user's detection result. Use user reliability to remove malicious users in each cell. Calculate the accumulated average value of reliable decision results of each cell and then compute location reliability. Finally, fusion center calculate a weighted objective function to make the final decision. According to simulation results, when cognitive users are mobile, this proposed mobile scenario algorithm can resist attacks from malicious users effectively, i.e. on condition that the false alarm probability is constant, it can increase the detection probability effectively.