Abstract: Underwater images typically exhibit low contrast, color distortion, and detail loss due to light scattering and absorption. To mitigate these issues, a novel underwater image visual enhancement algorithm based on a gray space prior is proposed. The algorithm first applies an improved Gaussian model for image preprocessing, addressing the color shifts caused by the underwater environment and restoring the natural perceived colors of the image. The balanced image is then mapped from the RGB space to the gray space, where the dark channel prior algorithm estimates the transmission map and ambient light values in the gray space. These estimates, combined with the underwater imaging model, are used to compute the noise characteristic distribution. Finally, the gray space prior is treated as a mapping hub, correlating features in the color and gray spaces to eliminate visual blurriness in the RGB space caused by noise, and improving the perceived clarity of the degraded underwater image. Experimental results on publicly available underwater benchmark datasets demonstrate that this approach offers superior performance in color balance and detail enhancement compared with existing underwater image enhancement methods.