Landscape is a product of the interaction between the human and natural components of our environment, reflecting the hierarchical structures and complex interrelationships of natural ecosystems and socio-cultural dimensions across multiple spatial and temporal scales. Landscape character assessment (LCA), the process of identifying and describing variation in the character of the landscape, enabled the identification of distinct, recognizable, and consistent patterns in landscapes. As such, it was regarded as a fundamental procedure for the sustainable planning and management of landscape resources. The region along the Yellow River in Inner Mongolia was situated in an arid-semi-arid transitional zone, where the complex natural environment and the multicultural integration of agriculture and pastoralism have collectively shaped a unique regional landscape. In this study, grounded in social-ecological systems (SES) theory, we developed an integrated framework for landscape character identification from a systemic perspective that incorporated ecological, social, and human-nature interactions. Taking the five cities along the Yellow River in Inner Mongolia as a case study, we constructed a comprehensive indicator system comprising 24 distinct elements across six dimensions: nature and ecology, socio-economic development, cultural and recreation, land use patterns, spatial heterogeneity, and temporal dynamics. The indicators synthesized multi-source datasets, including land use data (1980-2020), Points of Interest (POI) data, and Historical Geographic Information System (HGIS) data, et al., and combined with multivariate spatial analysis implemented through GIS. The Self-Organizing Map (SOM), an artificial neural network model employing unsupervised learning, was used to identify and classify landscape character types (LCT). It offered efficient data processing capabilities for handling high-dimensional data and complex relationships. Through this methodology, 24 landscape character types were identified within the research area, followed by visualization, characteristic description, and spatial pattern analysis of the landscape classification. The results showed that : (1) Natural and semi-natural landscapes dominate the region; rich natural-cultural composite landscape resources formed a linear heritage corridor along the Great Wall, spanning the northern plateau. (2) Hohhot and Baotou exhibited significant clustering in both high socio-economic landscapes and high cultural-recreational landscapes. (3) Complex human-land coupling relationships existed in the Yellow River coastal plain and hilly-mountainous areas in the eastern part of the study area, where the spatial heterogeneity of the landscape and its distinct east-west and north-south differentiation in these areas were caused by a combination of human activities and the natural environment. This study proposed a framework and methodology that effectively identified landscape character types with multi-dimensional attributes, offering insights into the complexity of human-nature interactions and the diverse functions of landscapes. It provides support for multi-objective landscape planning and management at the regional scale, including sustainable development, ecological restoration, and cultural heritage conservation.