The enhancement of ecosystem function within the Hohhot-Baotou-Ordos-Yulin urban agglomeration is significant for advancing ecological protection and promoting high-quality development in the Yellow River basin. Ecosystem service value (ESV) was evaluated based on the land use types in the study area from 1990 to 2020, the random forest model coupled with the SHAP method were used to analyze the importance, response, and interaction of the driving factors, and partial least squares path modeling was used to identify interaction paths of the driving factors. The results showed that (1) the area of cropland, grassland, and water exhibited an initial increase followed by a decrease, while the construction land area continued to increase. The ESV underwent an initial decline, followed by an increase of 62.28 billion yuan, primarily due to the conversion of cropland and grassland to forest. (2) Land use was the most important driving factor of ESV change, with a 61.24% contribution rate. Natural factors, including topography and meteorology, were the next significant contributors, with 17.59% and 17.05% contribution rates, respectively. Socio-economics factors had the least impact on ESV, with only 1.73% contribution rate. The responses of ESV to driving factors were characterized by nonlinear changes. (3) The proportion of water was the most important interaction factor, and the interaction between driving factors showed different effects when the driving factors were in different ranges. Topography, meteorology, soil, and socio-economics indirectly influenced ESV by affecting land use. The direct effects of socio-economics on ESV were negative, while its indirect effects were positive. It recommended to strengthen the protection of grassland, forest, water, and other ecological spaces and further to coordinate their relationship with urban construction and mineral resource development. Furthermore, the implementation of multi-factor collaborative management and control of regional ecosystems was oriented towards improving service functions.