Ecological restoration in degraded landscapes requires a comprehensive understanding of the factors influencing ecosystem function, particularly in relation to water and carbon cycling. This study explores the role of microtopography and plant functional traits in optimizing water-carbon coupling efficiency in a mining-affected ecosystem using the CATS model. We assessed the water-carbon coupling index (WCCI) across five microhabitat zones (A–E) within a mining area in the Hulunbuir Grassland. Results show significant variability in WCCI across zones, with Zone B exhibiting the highest functional efficiency due to its moderate moisture and low erosion, while Zone A displayed the lowest WCCI, constrained by water and nutrient limitations. The CATS model simulations revealed that water-carbon coupling is highly influenced by species functional traits such as SLA, height, and drought tolerance, with species like polygonum aviculare and cleistogenes caespitosa contributing most significantly to functional performance. Additionally, ecological filters, such as soil moisture, nutrient availability, and erosion intensity, were found to shape species selection and community structure. Our findings highlight the importance of trait-based approaches in restoration, emphasizing the need for tailored species optimization that accounts for both functional trait diversity and local environmental conditions. This research offers valuable insights for improving ecosystem resilience and optimizing water-carbon coupling in the face of climate change and land degradation.