<p>Learning to rank is a widely used approach for ordering items based on user preferences, but it faces challenges due to uncertainty caused by ambiguous queries, diverse information sources, and inconsistent relevance labels. This systematic review explores techniques for managing uncertainty in LTR, identifying trends, and addressing challenges. The study highlights a variety of methods, with probabilistic models, such as Bayesian learning, variational inference, and Monte Carlo dropout, being the most prominent due to their rigorous mathematical foundations. Fuzzy-based techniques, including fuzzy set theory and fuzzy rule-based systems, offer flexibility in handling uncertainty, while evidence theory and fuzzy integrals provide unique but less commonly used perspectives. Deep learning models, such as Bayesian Neural Networks (BNNs) and Monte Carlo Dropout, have gained traction for uncertainty management in ranking tasks. Techniques like deep ensembles, attention mechanisms, and hybrid models enhance robustness, feature relevance, and interpretability, particularly in critical domains like healthcare and finance. Addressing uncertainty is essential for improving the accuracy, relevance, and robustness of ranking systems, which directly impacts user satisfaction and system performance. This review synthesizes existing techniques, highlighting their strengths and limitations, and provides actionable insights for researchers. By enabling the selection of appropriate uncertainty handling methods tailored to specific domains, the study contributes to the development of more effective ranking systems. The research draws from four major digital databases (Web of Science, Scopus, ScienceDirect, and Springer), categorizes techniques hierarchically, and offers a comparative analysis of methods.</p>