Precision agriculture relies on the uniformity of location-based agricultural environmental data, such as soil, weather, and sensing. In this study, we propose a grid system for precision agriculture to data spatial alignment. Based on this precision agriculture grid system, we analyzed the spatial distribution variability of annual soil chemical components and estimated rice yield distribution based on these components. The results from the years 2018 and 2019 showed that the average values of soil chemical components in non-paved areas exhibited no significant differences. However, even with the similarity in average soil chemical component values, we confirmed through spatial distribution maps that there can be diverse and biased distributions within the soil, as well as inter-annual variations. In the process of estimating rice yield using soil chemical components, the Random Forest model showed higher explanatory power and accuracy compared to the multivariate linear regression model. Both models, however, incorporated a wide range of chemical characteristics as model variables, beyond the key variables typically used in traditional rice fertilizer estimation. This diversity in model variables is presumed to be a result of the unique agricultural environment, such as uneven use of additional fertilizer. Therefore, it is crucial to emphasize that, even after calculating crop fertilizer quantities through soil analysis, stable yield and quality assurance require precision agriculture management based on crop growth monitoring.