Traditional image contrast enhancement technology can easily cause over-enhancement or under-enhancement in the local area of images. Thus, an image contrast enhancement method combining nonseparable wavelet analysis and curve fitting is proposed. The method uses mean and standard deviation of the standard clear image as the ideal target for image contrast enhancement. Firstly, four-channel nonseparable wavelet is used to decompose standard clear image and low contrast image of the same target. Next, the predictive ability of curve fitting is used to get the relationships among mean values and the relationships among standard deviations of sub-images with same frequency channel and same direction. Then, the ideal enhanced sub-images can be obtained by the functional relationships and histogram matching formulas. Finally, these new sub-images are reconstructed by nonseparable wavelet. The performance of the method is evaluated with clarity and standard deviation. The experimental results show that the proposed method has a good enhancement effect for low contrast image, and can obtain high definition images.