Resource allocation in dynamic maritime wireless mobile networks is essential for uninterrupted communication due to the complexities posed by the marine environment. Traditional methods of resource allocation still fail to address the challenges posed by bandwidth variability, mobility patterns, and coverage areas even with strong technological advancement. This research is concerned with the optimization of resource allocation processes in wireless mobile networks using machine learning (ML) algorithms to increase reliability, throughput, and reduce latency. Various machine learning algorithms including supervised learning and reinforcement learning were analysed using real and simulated datasets from the maritime domain. This study focuses on creating the experimental design for training and validating the models with the golden metrics of packet delivery ratio, spectral efficiency, and network utilization. Results show that ML resource allocation strategies automate and anticipate dynamic changes, allowing for better resource allocation adaptability, therefore significantly outperforming traditional techniques. The results indicate that there is great potential for improving the maritime communication infrastructure through intelligent solutions. This study serves as a guide and basic framework for future monitoring and research in maritime networking for real world applications.