Cognitive radio non-orthogonal multiple access (CR-NOMA) networks promise improved spectrum utilization and capacity in 5G networks. In this work, we aim to investigate efficient power allocation for the secondary users (SUs) in underlay CR-NOMA networks using a game-theoretic approach. We present a novel power allocation to CR-NOMA network from a game-theoretic perspective. First, we specify the utility function of the primary users (PUs) and SUs, and formulate the game as a non-cooperative game. Then, the existence and uniqueness of the Nash equilibrium (NE) are investigated. Finally, the sum utilities of SUs is maximized by optimal power allocation at the NE point. Simulation results provided that the proposed scheme outperforms the conventional method, providing up to 37.5\% increase in sum utilities of the SUs.