Climate is still the main responsible for soybean yield variations, exerting a limiting action on these agricultural systems. In the cerrado “savannah” of Bom Jesus, this crop has shown a growth in cultivated areas over the years, however, productivity does not follow the same rhythm, going through oscillation periods. Although cultivation is aggregated to high technology, the crop is highly vulnerable to climatic adversities. Thus, the objective of this study was to analyze possible trends in meteorological variables that could influence soybean yield in the municipality of Bom Jesus (Piauí). Daily meteorological data were used for the variables: precipitation, maximum and minimum temperature and air diurnal thermal amplitude from 1974 to 2014. The statistical Mann-Kendall test was applied in order to analyze variation trend in climatological time series of the variables under study. As regards variability of the agrometeorological elements, a significant negative trend for precipitation in the months of June and October and a significant positive trend in the month of December was observed from the indicated results, which detected changes in the local climate patterns. It was also possible to identify statistically significant positive trends in the maximum temperature, for all the months that are part of the soybean cycle (November - April). The identified results may be used to elaborate planning strategies to choose the best sowing time for soybean crop in the municipality of Bom Jesus (PI), in order to minimize the vulnerability of the crop productive system to climatic risks.