Vegetation cover is one of the most valuable environmental assets of Baghdad city, yet It has contracted sharply under urban pressure, informal settlement expansion, water scarcity and rising surface temperature. This paper reviews advances in remote sensing (RS), Geographic Information Systems (GIS) and machine learning (ML) for land cover / land use (LC/LU) classification, focusing on the measurement of vegetation shrinkage in Baghdad. Worldwide and Iraqi studies published between 2018 and 2026 were evaluated, covering multi-temporal Landsat and Sentinel imagery, spectral indices, supervised classifiers and post-classification change detection. Globally, research has moved from pixel-based statistical classifiers to ensemble learning, deep learning and hybrid predictive frameworks. Iraqi work is growing but still relies on medium-resolution imagery conventional classifiers and short observation windows. Persistent obstacles include inconsistent classification schemes, spectral confusion between sparse vegetation and bare soil incomplete accuracy Reporting, and the absence of a unified national land cover database for Iraq.