By means of the wireless transmission model of Rayleigh fading,people often use the utility function of cognitive users to solve the problem of channel programming in cognitive radio networks and to further obtain the transmission rate.As the final communication utility only relates to the water-level when both the total transmission power and the bit error rate（BER） of cognitive users are constrained,a algorithm of power allocation based on water-level adjustment is proposed,through which power and subcarriers can be assigned reasonably and the transmission rate can be finally obtained.Theoretical and simulative results show that the proposed algorithm is of low computation complexity and high utility close to that of the optimal algorithm,and that the adjustment factor only relates to the convergence rate of the algorithm,but not to the total utility and channel allocation results of cognitive users.