<p>Urban expansion in Europe has accelerated over recent decades, significantly blurring the boundaries between urban and rural landscapes and giving rise to extensive peri-urban areas. These dynamic transformations demand timely and effective landscape classifications to track environmental, demographic, and socio-economic developments and for a deeper understanding of human settlements and their sustainability impacts. Traditional threshold-based classification methods, while widely used, rely heavily on expert knowledge and require substantial human input, posing challenges for rapid updates to rural–urban classifications. This study addresses these limitations by leveraging an unsupervised learning neural network to classify territorial typologies in Switzerland at a granular municipal level. Using a comprehensive dataset that integrates physical, demographic, and socio-economic variables, we identified six distinct clusters, which were subsequently mapped onto the geographical space, providing a characterization of the Swiss landscape through its intricate rural–urban interface, urban expansion frontiers, and aging populations in peri-urban and rural areas. Our analysis also highlights critical phenomena such as agricultural land loss and the encroachment into ecologically sensitive Alpine territories. Graphical outputs, such as heatmaps and boxplots, were used as effective visual tools to aid in results interpretation, uncover variables distributions and relationships, and enhance the analytical process. The findings demonstrate that clustering procedures based on unsupervised machine learning provide valuable insights into rural–urban configurations and contemporary urbanization processes, offering a robust framework for studying the interplay between human settlements and environmental sustainability, particularly in vulnerable Alpine regions.</p>