Focused on the problem that noise is enhanced with image enhancement in the traditional image enhancement methods an image enhancement algorithm based on wavelet transform and improved fuzzy set theory was presented. Firstly, the multi-scale wavelet transform was adopted to decompose the input image. Secondly, the low frequency coefficients were enhanced by the linear piecewise function and wavelet threshold was used for the high frequency coefficients de-noising, then a new membership function of fuzzy was defined and the high frequency coefficients of different directions of each scale were enhanced by fuzzy enhancement transformation. Finally, the inverse wavelet transform was applied to synthesis image. A group of experimental results demonstrate that the disadvantages of traditional enhancement methods are avoided, and the presented algorithm can enhance the key characteristic of the image and restrain noise effectively.