Abstract: In response to the poor adaptability of network routing in complex and dynamic marine environments for underwater acoustic sensor network (UASN), which leads to significant fluctuations in network performance and low robustness, a comprehensive link quality cognitive routing protocol (CLQ-CR) is proposed. This protocol evaluates link quality and optimizes routing through environmental cognition at different levels and mechanisms. Firstly, the comprehensive link quality factor, such as the K-factor, signal-to-noise ratio, and node load, is defined to comprehensively qualify the UASN link from the perspective of signal, channel, and link feature in the marine environment. Low-cost link quality assessment is achieved by combining multiple access contention and monitoring. Based on this, a multi-attribute optimization mechanism is designed using the technique for order preference by similarity to ideal solution (TOPSIS) for cognitive optimization of UASN routing. The simulation and experimental results of UASN conducted in practical shallow sea environment show that the proposed method can achieve better network performance indicators such as packet delivery rate than fixed routing and conventional cognitive routing methods, thus verifying the effectiveness of improving the adaptability of underwater acoustic sensing networks to marine environments through cognitive optimization.