Abstract: In order to increase the energy efficiency of secondary users in cognitive radio networks, the processes of spectrum sensing and access of secondary users are modeled by continuous-time Markov theory. Based on the analysis of the main factors affecting the energy efficiency of secondary users, an energy-efficient algorithm based on cross-layer design is proposed. The algorithm can decrease communication interruption probability of secondary users caused by unslotted access of primary users, and a tradeoff between sensing performance and energy effi-ciency is gained by setting proper sensing time and access probability. Simulation results indicate that the proposed algorithm is better than single-layer optimal schemes which only consider either spectrum sensing or access strategy for it provides the secondary users with more energy efficiency.