Land degradation is a global environmental issue that affects the ecological functions and sustainability of land resources. In Indonesia, degraded land conditions are commonly addressed through the concept of critical land, particularly in watershed assessment, management, and rehabilitation. Spatial approaches have been widely applied to identify critical land over large areas; however, their representativeness is strongly influenced by input data quality and may not fully reflect site-level degradation conditions. This study assessed critical land using a spatial approach that integrated key biophysical parameters with field-based verification in part of the Serang Watershed. Critical land was identified using a multicriteria spatial approach based on P.3/PDASHL/SET/KUM.1/7/2018. Land cover was derived from Sentinel-2A imagery using Random Forest and quantitatively assessed using 45 field reference points. Soil erosion was estimated using USLE, with soil erodibility determined through laboratory analysis of 18 soil samples representing different land units. Five criticality classes were identified, with Slightly Critical land dominating 58.45% of the study area, while Critical and Highly Critical land accounted for 6.27%. Field-based verification of areas classified as Critical and Highly Critical revealed diverse degradation characteristics, including rill and gully erosion, topsoil loss, shallow soil, slope failure, and anthropogenic disturbances. These findings indicate that integrating spatial modelling with field-based verification can provide site-level information to support the interpretation of critical-land conditions and the development of site-specific soil conservation and land rehabilitation measures.