Accurate and reliable agricultural forecasting faces increasing challenges amid growing environmental variability. Data assimilation (DA), which coherently integrates observational data within process-based models, offers a powerful solution to improve real-time crop monitoring, anomalies detection, and dynamically updated predictions. By narrowing the gap between modeled projections and on-the-ground realities, DA enables more informed decision-making for farmers, policymakers, and researchers alike. However, care must be taken to prevent runaway errors or implausible simulation conditions.This session invites submissions highlighting recent advances in data assimilation for agricultural applications, including—but not limited to: innovative DA methodologies for crop and land surface models; use of novel datasets including remote sensing, ground observations and agricultural management surveys; nexus of machine learning and DA; multi-scale, multi-source and multi-variable assimilation while maintaining biophysical balance; operational implementation for decision support.