Yield forecasting has been extensively tackled by the use of crop growth models, however its implementation at field scale still faces numerous constrains due to the lack of input data required and uncertainties in parameter value. Current accessibility of remote sensing data with global coverage and free access offers a great opportunity to enhance crop model predictions throughout retrieval biophysical parameters to link with model, such as leaf are index (LAI). Among methods available to assimilate LAI into the model, calibration has been successfully applied for wheat (Gaso, et al., 2019). Therefore, the aim of this study was to predict spatial variability of soybean grain yield and maximum root depth (RDmax) at field scale by assimilating temporal series of Sentinel-2.