Land cover mapping is the process of creating a map that shows the types of land cover in a given area. This includes vegetation, land use, surface features, and other natural and human-made features. Land cover mapping is used to help understand environmental processes, land use, and land management. It also helps identify potential risks from natural disasters, such as floods and landslides. By understanding land cover, land managers can make decisions about how to best use and manage the land. Deep learning has opened up new opportunities for the extraction of LC information from remotely sensed imagery.. Deep learning models are typically applied to satellite imagery and aerial photographs, which are then analyzed with the help of convolutional neural networks (CNNs). CNNs are able to accurately identify and classify land cover types in high-resolution imagery. In this paper we have implemented 3 Semantic Segmentation models namely PSPNet, UNet and FPN on the DeepGlobe dataset for land cover mapping. We have achieved IoU scores of 91%, 78% and 64% respectively.