The sports industry emphasizes injury prevention, but current methods struggle due to data complexity. Machine learning (ML) offers promise by synthesizing athlete data for informed decisions. Research proposes ML-based decision support systems, simplifying variables for better insights. Classification algorithms dominate injury prediction, with ML showing promise but lacking consistency in predictors. Short-term models aid coaching staff, while long-term models benefit performance and operational departments. Challenges include addressing temporal dependencies and data leakage during preprocessing. Thorough data inspection and domain knowledge are crucial to prevent overconfidence in model predictions. ML presents an opportunity to advance injury prediction and athlete management in sports, but it’s important to address key challenges and ensure robustness before using them in practice.