de Oliveira Aparecido, Lucas EduardoTorsoni, Guilherme Botegada Silva Cabral de Moraes, José Reinaldode Meneses, Kamila Cunha [UNESP]Lorençone, João AntonioLorençone, Pedro Antonio2020-12-122020-12-122020-01-01Environment, Development and Sustainability.1573-29751387-585Xhttp://hdl.handle.net/11449/198977The study of the soybean yield variability influenced by the climate contributes to the planning of strategies to mitigate its negative effects. Thus, our aim was to calibrate agrometeorological models for soybean yield forecast and identify the weather variables that most influence soybean yield. This study used historical series of climate and soybean yield data from soybean-producing locations in the Mato Grosso do Sul state, Brazil. The historical climate series was 20 years (2000–2019). The soybean production, yield, and planted area data of the localities were in the period from 2009–2018. Multiple linear regression analysis was the statistical tool used for data modeling. The models from the north and central regions forecast of anticipation of 2 months since the final data necessary to apply the model were EXCJANc and PJANc, respectively. The models calibrated for the southern region reported anticipation of one month since the final data necessary to apply the model was EXCFEVc. The calibrated models used to forecast soybean yield as a function of climatic conditions have a high degree of significance (p < 0.05), high accuracy and errors lower. The models for the northern and central regions show a prevision of anticipation of 2 months before soybean harvest, a period that is essential for producers to be able to conduct pre- and post-harvest planning. The climate variable with the greatest negative influence (r = − 0.54) on soybean yield in Mato Grosso do Sul state was water stress in December.engClimateCrop modelingGlycine max LSpatial error modelYield zoningArtigo10.1007/s10668-020-00807-w2-s2.0-85086386676