In order to realize the potential of link adaptation, reliable channel prediction is necessary. In this paper, we propose a novel channel predictor based on Constrained Hidden Markov Model (CHMM). By partitioning the range of the received signal envelope into several intervals, a CHMM can be constructed with the high efficiency algorithm. Then an improved prediction method is presented, which is more accurate than the simple prediction method of the largest transition probability. Finally, simulation results are given to show the effectiveness of the CHMM channel predictor. In this paper, we propose a novel channel predictor based on Constrained Hidden Markov Model (CHMM). By partitioning the range of the received signal envelope into several intervals, a Then, an improved prediction method is presented, which is more accurate than the simple prediction method of the largest transition probability. Finally, the simulation results are given to show the effectiveness of the CHMM channel predictor.