In the internet of things IoT applications, cooperative spectrum sensing of cognitive radio sensor networks CRSN may use electronic objects or sensor nodes to cooperatively detect the spectrum of primary user. How to effectively fuse the local detection data and make a global decision is critical for CRSN. In order to improve the detection accuracy of CRSN, we propose a cooperative spectrum sensing scheme based on side information, and the scheme uses a cooperative spectrum sensing framework and an efficient clustering algorithm that can meet the specific requirements of IoT applications. Though mathematical modelling, minimising the missing detection probability is converted into clustering nodes. The clustering algorithm is to be used to find the optimal distance and localisation of nodes. Simulation results show that the proposed cooperative spectrum sensing scheme has better performance than the conventional cognitive radio models such as equal gain combining, and maximal ratio combining algorithm. In addition, the influence of sensor nodes' density and distance range on the sensing performance is also analysed.