Abstract: The current mobile communication system has problems such as low spectrum utilization, high energy consumption, and low energy efficiency. In the Cognitive Radio Non-Orthogonal Multiple Access (CR-NOMA) network system, a power allocation scheme based on improved artificial fish swarm algorithm is proposed. The system includes multiple primary and secondary users. Firstly, in order to improve spectral efficiency, reduce error probability and delay of demodulation, secondary users access the system in the form of non-orthogonal multiple access, and a user grouping strategy with uniform channel gain difference is adopted. Secondly, considering that the traditional artificial fish swarm algorithm is easy to fall into the local optimal solution, weak optimization ability and poor population diversity for the secondary user power optimization, the constraint operator mechanism and adaptive strategy are introduced into the artificial fish swarm algorithm. Finally, the algorithm is used to jointly optimize the power between each subchannel and the power of secondary users within the subchannel, seeking the optimal transmission power of secondary users, and maximizing the total energy efficiency of the system. The experimental results show that when the secondary user is 30, compared with the traditional artificial fish swarm algorithm, the total energy efficiency of improved artificial fish swarm algorithm is improved by 10.6% , and the system performance is better.