Abstract: A power allocation algorithm supporting delay constrains which maximizes the channel effective capacity of secondary users for satellite "Underlay"cognitive radio networks is proposed to improve the spectrum utilization of a satellite communication system. A power interference model is established based on the topology of satellite cognitive radio networks first,and then the object functions of effective capacity are derived both in perfect and imperfect channel state information environments. The Lagrange method is used to obtain the optimized transmitting power of secondary users by introducing the time-domain channel correlation coefficient.Optimized power adjusting strategies under different scenes are discussed while the power allocating and optimizing process is simplified. Factors which affect the effective capacity of secondary users are analyzed by simulation. The results indicate that the proposed power allocation algorithm can adjust the transmitting power of secondary users according to delay constrains with channel fading characteristics,and the channel effective capacity of secondary users is improved obviously compared to the traditional equal power allocation policy.