<p>Coal mining subsidence in areas with high groundwater levels can cause surface subsidence, leading to the formation of water bodies that inundate agricultural land, crops, and residential areas. These conditions present significant ecological and environmental challenges. Precise identification and classification of waterlogged regions are essential for effective management and ecological reclamation. This study developed a comprehensive method for identifying and classifying water bodies in mining areas, focusing on subsidence-induced water bodies. Firstly, the study applied an enhanced modified normalized difference water index (MNDWI) for feature extraction, supported by edge detection, morphological post-processing, and OTSU threshold segmentation to ensure precision. Next, we used long-term Landsat imagery (1984–2022) to extract water features, monitor temporal changes, and distinguish between natural and subsidence-induced water bodies. Additionally, several similarity measures, including Hausdorff distance, Fréchet distance, and DTW, were used to enhance classification accuracy. Finally, a voting-based mechanism integrates results from multiple algorithms to ensure robust and reliable classification. The key findings are as follows. (1) The MNDWI-based approach identifies water features more effectively than the traditional NDWI, offering high efficiency without requiring training data, making it ideal for long-term water extraction. (2) Temporal analysis shows that subsidence-induced water bodies display distinct growth patterns compared to natural ones, with areas gradually increasing and stabilizing over time. (3) The proposed method reliably differentiates between subsidence-induced and natural water bodies, validated through similarity metrics and a voting mechanism, ensuring accurate classification across the study area. This study provides valuable insights for environmental management in mining regions.</p>