This thesis proposes a routing algorithm to enhance throughput, spectrum-efficient communication and to fulfill environment-friendly for underwater cognitive acoustic sensor networks (UCSNs). In a UCSN, each cognitive acoustic (CA) user would sense the medium to detect available spectrum resources. However, CA users have no exact knowledge about the communication mechanism of the primary users (PUs) such as sonar users or marine mammals. In addition, because of the frequency-dependent attenuation, the available frequencies in water are severely limited and are still underutilized. Therefore, by analyzing the behavior of the CA users based on the ON-OFF process in the primary channel, an efficient routing algorithm for UCSNs is proposed to enhance throughput while limiting interference to PUs to establish friendly communication. Finally, simulation results are shown to verify the effectiveness of the proposed scheme.