Process-based models can simulate the interactive effects of management practices and weather on key agroecosystem outcomes. Utilizing modeling tools to improve ecosystem productivity, soil health, and resource-use efficiency in croplands, grasslands, and agroforestry systems is critical but remains challenging because of structural, parametric, and observational uncertainties. This session invites studies that advance process-based modeling for agricultural applications through improved farm practice evaluation and decision support. We welcome contributions from the field to global scales that address (1) model intercomparison and ensemble for improved assessment of management practices across regions, (2) simultaneous evaluation of multiple agroecosystem outcomes to quantify co-benefits/tradeoffs, (3) the use of AI/ML or large language models to translate complex mechanistic model outputs into decision-ready formats, and (4) robust comparison between regenerative and conservation practices through dynamic baseline methods. Studies that incorporate remote sensing, data assimilation, stakeholder co-development, and other novel aspects of process-based models are also encouraged.