Abstract: In this paper the optimum design of an adaptive modulation based on uncoded MQAM modulation assisted by wireless channel prediction for the flat Rayleigh fading chan-nel is discussed. By studying how a known prediction error variance will affect the optimized transmission properties, optimum solutions for the rate are derived based on the predicted SNR and the prediction error variance. And the data rate is adapted to maximize the spectral efficiency subject to average bit error rate constraints. The numerical results by simulations show that it adjusts the link adaptation better to the true channel conditions and improves the overall system performance.