Impacts due to climate change, population growth and intensive agriculture continue to be a major concern worldwide. Sustainable agriculture with coherent land management strategies is essential to mitigate against adverse environmental impacts. For the Chinese Loess Plateau (CLP), much research has focused on implementing soil-plant-atmosphere models to inform mitigation initiatives such as large-scale vegetation restoration programs. However, model choice typically depends on measurement availability and specific research questions, where many modeling approaches are established according to site specific data and parameterized via local information, making their generalization elsewhere difficult. Furthermore, in most studies only one modeling approach is selected, and thus its merit is difficult to assess relative to alternatives. Given these challenges, this review examines the capability of models with different level of complexity to simulate water fluxes and nutrient transformations for the CLP. Reviewed models were typically employed under different climate conditions (e.g., snowmelt, soil freezing and thawing) and across different land-uses (e.g., revegetated areas) which reflects the robustness of some models (e.g., for description of vegetation grow), but at the same time illustrates model weaknesses that should be addressed (e.g., water simulations under thawing conditions). On conducting this review, a general framework for choosing or developing the most appropriate modeling approach is established given a study site's climatic and ecological conditions and research aims.