Underwater images gain popularity in different applications, namely, marine engineering and exploring ocean resources. However, these images face hazing and color degradation effects due to improper light and absorption of color. To overcome these limitations, several underwater image enhancement models have been proposed in the literature. The existing models enhance the images at the same level without considering the image characteristics. However, in the real world, characteristics of the image vary for different images. Therefore, in this paper, we have designed an underwater image enhancement model based on the swarm intelligence termite alate optimization (TAO) algorithm. This algorithm determines the optimal parameters of the enhancement methods based on the image characteristics. The TAO algorithm provides better exploration, exploitation rate, and required minimum parameters over other optimization algorithms. Thus, it is chosen for the proposed model. The proposed model is categorized into two phases. In the first phase, the underwater image is read. Further, its RGB channels are extracted. The RGB channels are classified into three planes, namely, superior, inferior, and intermediate, based on the mean value of the image pixels. After that, the power law method is applied to the superior plane because this plane is used as a reference plane to enhance the other planes. In the second phase, the singular value decomposition (SVD) method is applied to inferior and intermediate planes based on superior plane characteristics. The optimal gamma and scaling values of the power law and SVD method are determined using the TAO algorithm. Besides that, exposure value and entropy are taken as the objective functions in these methods. Further, underwater standard dataset images are taken into consideration to validate the performance of the proposed underwater image enhancement model. Next, visual quality is checked using subjective analysis, and image characteristics are analyzed based on objective analysis. Finally, a comparative analysis is performed based on various parameters. The results show that the proposed model achieves better enhancement in terms of visual quality and entropy parameters.