A novel variational blind restoration approach based on the sparse prior of red channel is proposed according to the complete underwater optical image formation model. In order to simultaneously tackle the problems such as haze, low contrast, color distortion, and blur caused by the scattering and absorption of water and suspended particles in underwater scene, multiple regular terms with different purposes are merged into the proposed variational model. First, a guided image is produced for color rendering depending on the direct component and backscattering component of underwater optical imaging model. Subsequently, the data fidelity term is designed based on the forward scattering component. Additional, L0 norm form is introduced as the deblurring regular term on the basis of the sparse prior of red channel. Moreover, to accelerate the computational efficiency, an alternating direction multiplier method is employed to solve the proposed model. Experimental results demonstrate that the proposed method not only can remove haze, enhance contrast, and recovery color, but also has a good performance on deblurring and improving visibility.