Abstract: In both the academic and engineering fields, the pursuit of obtaining color images with certain availability in severely bandwidth-constrained underwater acoustic channels is a matter of great interest and attention. In this paper, a new method for underwater color image transmission is proposed, which utilizes the set partitioning in hierarchical trees (SPIHT) algorithm based progressive image transmission and visual salience detection to achieve underwater color images with better availability in complex and highly bandwidth-constrained underwater acoustic channels. The method dynamically adjusts the data transmission scheme based on the signal-to-noise ratio and employs red channel compensation to enhance the accuracy of saliency detection in the frequency domain. Subsequently, the SPIHT is used for progressive image transmission, and the block artifacts caused by high down sampling rates are addressed by using guided filtering at the receiver end to obtain high-quality underwater images. Experimental results demonstrate the applicability of the proposed method in compressing underwater color images.