Regional land use/cover change and future development scenarios are of great significance for the regional sustainable development and land management. In this study, we selected northern Shanxi province, an agro-pastoral ecotone, as the research area. The land use/land cover (LULC) maps of the study area in 2010 and 2015, as well as the driving factors including elevation, population, economy, temperature, and precipitation were extracted from various national-level datasets. We first used the CLUE-S model to simulate the LULC of 2015 based on 2010 LULC data and driving factors, and evaluated the simulation accuracy with the obtained 2015 LULC data. Then, we used the evaluated CLUE-S model to simulate the LULC patterns in 2020 under three future development scenarios. The results showed that the LULC of northern Shanxi was dominated by cultivated land, forestland, and grassland. The LULCs mainly distributed in strip-shapes toward the northwest. The logistic regression model could extract the relationship between LULC and driving factors well, and reflected the effect and degree of different driving factors on the LULC pattern. The CLUE-S model had high accuracy in fitting the LULC of northern Shanxi, with a Kappa coefficient of 0.89. The scenario simulation showed that the LULC under the environmental protection scenario (c) was more sustainable than the business-as-usual scenario (a) and economic development scenario (b) in the study area. The results suggest that the increasing rate of the industrial and mining land should be slowed down, and the size of construction land should be strictly controlled in future land development, and the LULC pattern should be further optimized.