Abstract: To comprehensively evaluate the ecological function characteristics of artificial forests of the loess gully region in northern Shanxi and to identify their driving factors, this study examined pure and mixed stands of Pinus tabulaeformis, Pinus sylvestris var. mongolica, and Caragana korshinskii. A multi-criteria evaluation system comprising 15 indicators across five criterion layers—water conservation, soil conservation, carbon sequestration and oxygen release, biodiversity protection, and stand productivity—was established. Principal component analysis (PCA) and cluster analysis methods were applied to quantify the ecological function status and levels of different stand types, while the geographic detector model was used to explore the driving mechanisms of spatial differentiation in the ecological functions of artificial forests. The results showed that mixed forests exhibited significantly higher ecological function than pure stands, with Pinus tabulaeformis-Caragana korshinskii mixed forests performing best. Ecological function levels of artificial forests displayed clear stratification, dominated by good and medium grade plots. Forest type and NDVI were identified as the primary factors controlling spatial variation in ecological function. Most interactions between factors indicated enhancement effects, suggesting that ecological functions were jointly regulated by vegetation attributes, topography and hydrology, and stand structure. The findings highlighted the importance of mixed afforestation and stand structural optimization to enhance the comprehensive ecological functions of artificial forests, providing a scientific basis for the management of artificial forests in fragile ecosystems of the Loess Plateau.