The surge in containerised trade has intensified the need for efficient resource allocation in port operations, particularly in berth allocation (BAP), quay crane assignment (QCAP), and quay crane scheduling (QCSP) problems. While mathematical programming and metaheuristic approaches have traditionally been used to solve these problems, their scalability and adaptability remain limited. Recent advances in Machine Learning (ML) offer new optimisation possibilities. This paper conducts a systematic literature review of ML-based approaches applied to BAP, QCAP, QCSP, and their integrated variants, following PRISMA guidelines and classifying the selected studies across nine analytical dimensions. The review highlights a predominant focus on economic optimisation, with BAP being the most frequently addressed problem. Earlier studies broadly applied regression techniques and predictive approaches. At the same time, more recent research has shifted towards deep reinforcement learning and prescriptive solutions, opening new opportunities for sustainable and adaptive port operations. Finally, it has identified key trends and opportunities in ML applications for port operations.