Abstract: Cooperative cognitive radio networks (CCRNs) are attracting more and more attentions for its high spectrum efficiency. When the channel state between the primary users (PUs) is worse in the CCRNs, PUs can give out a dedicated portion of channel resource to some secondary users (SUs) in exchange for the SUs’ helping them to complete their communications. Suppose the SUs can perfectly know the channel state information (CSI) between them and the PUs, and the SU source node is equipped with multiple antennas, beamforming technologies can be employed to provide SU the capability of accomplishing its own signal transmission and the assisting the PUs to transmit their messages at the same time without interfering with each other. By this way, the spectrum utilization efficiency is greatly improved. In this work, we design the beamforming parameters based on zero-forcing (ZF) and minimizing sum of weighted mean square errors (MSWMSE) at the source SU node, respectively. Simulation results show that both criteria can provide the co-channel interference (CCI) cancellation, and the MSWMSE has a much better performance in the average sum of mean square errors(ASMSE) of the PUs and SUs because it can optimally balance CCI and noise at the destinations. Moreover, different performance of the PUs link can be achieved by adjusting the weight factor.