Rapid urbanization in developing countries, including Iran, has placed considerable pressure on natural ecosystems and environmental resources. Zanjan, as one of the major cities in northwestern Iran, has experienced significant physical expansion in recent decades. The objective of this study is to examine the physical development of Zanjan City between 2000 and 2025 using satellite imagery and advanced machine learning techniques. In this research, Landsat 5, 8, and 9 satellite images with less than five percent cloud cover were employed. To improve land-use discrimination, false color composite images were generated, enabling the differentiation of vegetation, barren land, and built-up areas. For land-use classification, three machine learning algorithms were applied: k-nearest neighbor (KNN), maximum likelihood classification (MLC), and support vector machine (SVM). KNN operates based on similarity and distance, MLC relies on Bayesian probability theory, and SVM utilizes kernel functions to separate complex data effectively. The results revealed that the built-up area of Zanjan increased from approximately 26.8–28.4 km² in 2000 to about 27.4–35.3 km² in 2025. This expansion was accompanied by a relative decline in vegetation cover and changes in barren land, indicating a clear trend of converting natural and agricultural lands into urban areas. Accuracy assessment demonstrated that SVM outperformed KNN and MLC, providing higher classification accuracy. Overall, the findings of this study are consistent with similar research in Iran, which highlights the spatial growth of cities. The results suggest that machine learning techniques, particularly SVM, are effective tools for monitoring urban physical changes. Moreover, this study provides valuable insights for sustainable urban planning and the protection of natural resources in Zanjan.