Finding practical solutions to balance food production and environmental conservation is a grand challenge the world faces today. Extensive soybean expansion in South America, at the expense of natural vegetation, pasture and other cropland, is at the forefront of this challenge. We propose to integrate satellite observations, field measurements, economic models and an agricultural ecosystem model to quantify historical patterns of soybean expansion across South America, analyze the underlying socioeconomic drivers and project future trends. We will use Landsat, MODIS and Sentinel 2 data to characterize soybean cover and map soybean expansion from 2000 to present with continentally distributed ground data for training and validation. To understand individual landowners’ land-use decision, we will develop a spatially explicit econometric Land-Use Change (eLUC) model and use the satellite-derived soybean expansion map and other remote sensing-based land-use data to empirically calibrate the model. We will extend the Multi-Regional Input-Output analysis with soybean land footprint accounting to track the land resources embodied in the international soybean supply chain. By integrating these different observations and models, the main goal of our proposal is to improve our scientific understanding of land-cover and land-use change in South America.