The heavy metal pollution in water and soil has become a pressing environmental issue, posing risks to ecosystems and human health. Traditional monitoring methods are often costly and time-consuming, leading to the adopting of data science techniques for more efficient and accurate assessments. This review examines recent machine learning and predictive modelling studies to analyse heavy metal contamination. Machine learning models have been applied to predict contamination levels, assess environmental risk factors, and identify high-risk pollution zones. Additionally, remote sensing and probabilistic models have improved large-scale monitoring capabilities. While these advancements enhance contamination forecasting, challenges remain regarding model interpretability, computational efficiency, and data quality. Future research should focus on hybrid AI models, real-time monitoring systems, and standardised performance benchmarks to improve prediction accuracy and environmental decision-making. By integrating data science with environmental management, this study highlights the potential of predictive analytics in mitigating heavy metal pollution and fostering sustainable development.