In this paper, a new self adaptive contrast enhancement algorithm for infrared images is presented based on plateau histogram equalization. By self adaptively selecting threshold value, this algorithm can enhance objects efficiently , and constrain the background and noises greatly in infrared images. With the proposed method, the disadvantages of histogram equalization are overcome, and large computational amount decreased greatly compared to other plateau histogram equalization algorithm while achieving high contrast. Theory analysis and experiments show that it is an effective enhancement algorithm for infared images.