Aimed at the mistiness and color distortion of underwater optical colorful image, an underwater image restoration method based on transmittance optimization and color temperature adjustment is proposed. Dark channel prior algorithm is applied to the underwater optical imaging model to realize the image visualization. Transmittance is improved by guided filter to increase its estimation precision. Then the image should be divided to bright areas and normal areas by setting threshold. Adaptive restored image is got by transmittance optimization. Finally the color of image is restored by adjusting the position of gray axis to rectify color temperature. Experimental result shows that the proposed algorithm can not only improve the clearness of image largely, but also correct the color accurately. Compared to other algorithms, the proposed algorithm have good results and robustness.