Underwater Wireless Sensor Networks (UWSNs) play a critical role in real-time environmental monitoring during disasters such as earthquakes, tsunamis, and underwater landslides. However, highly dynamic and hazardous underwater conditions significantly degrade the performance of conventional routing protocols. This study proposes a Q-learning-based Adaptive Data Transfer Algorithm (QL-ADTA) to address data transmission challenges in disaster-prone UWSNs. The proposed approach utilizes a multi-objective reward function that incorporates key contextual parameters, including data priority, link quality, node survivability, debris density, and water current strength. The algorithm is evaluated using disaster simulated synthetic datasets and compared with conventional routing protocols such as Dynamic Source Routing (DSR) and AODV (Ad-hoc On-Demand Distance Vector). Experimental results show that QL-ADTA improves packet delivery ratio by 25%, reduces latency by 23%, and enhances energy efficiency by 20%.These results demonstrate the effectiveness of the proposed approach in improving communication reliability and network performance under challenging underwater conditions.