Land cover segmentation plays a pivotal role in various remote sensing applications. This research delves into the comparative analysis of different advanced segmentation models, specifically focusing on the UNet framework, UNet with backbones, YOLOv8, and Detectron2. The study aims to provide in-depth knowledge of the performance variations among these models concerning their accuracy, robustness, and computational efficiency in land cover segmentation tasks. A systematic evaluation employing performance measures such as Intersection over Union (IoU) is used to quantify the segmentation quality of each model. This comparative study not only highlights the strengths and weaknesses of each approach but also assists practitioners and researchers in selecting the most suitable model for specific land cover analysis requirements. The investigation involves a diverse set of remote sensing datasets, ensuring the generalization of the findings across different geographical regions and land cover types. By training and validating the models on these datasets, followed by a rigorous assessment of their segmentation performance, the study provides valuable insights into the field of land cover segmentation. The outcomes are expected to aid in advancing remote sensing applications and facilitating informed decision-making in land management and environmental studies. Additionally, the findings can guide further refinement of segmentation models, pushing the boundaries of their capabilities for more accuracy and efficiency in real-world land cover classification.