In this paper, having compared the image haze-removal algorithms for image restoration and image enhancement, the principle of image haze-removal algorithm using dark channel prior is analyzed and discussed. The shortcomings of the algorithm are pointed out. At the same time, image haze-removal algorithm using the dark-channel prior based on wavelet decomposition is proposed. There are three steps included: three channel wavelet transform for the image, image haze-removal algorithm using the dark-channel prior of the low frequency components for dark channel prior and upgrading sharpening for high frequency components. The experimental results show that the algorithm can improve the resolution of image after image haze-removal and have better performance in detail, but also can effectively reduce the running time, improve the operating speed.