When detecting man made targets hidden in foliage in a FOPEN UWBSAR image, trunk clutter is the primary disturbance. Whether the detection can be performed successfully or not is dependent on the performance of the trunk clutter suppression. Based on Hidden Markov Modeling, a new trunk clutter recognition and suppression method is proposed. From the experiments results, the recognition method proposed in this paper can recognize the trunk clutter in UWBSAR image effectively, and improve the target detection performance in foliage.