Relevance. The topic of precision farming based on the use of geoinformation technologies is developing every year in new approaches to applied research. Increasing the efficiency of agricultural production is a key task of the state policy of the Russian Federation. This study analyzes such a soil indicator as hardness, and on the basis of experimental work, the effect of this indicator on the yield of such an agricultural crop as maize is revealed. In 2024, compared with 2001, an expansion of maize sown areas by 4.0 times (by 1,960.5 thousand hectares) is noted. Aim. Application of geoinformation technologies and machine learning to assess the possible impact of soil hardness on maize yield. Objects. 3D geospatial data in precision agriculture. Methods. Clustering, machine learning, ensemble, bayesian optimization methods. Results. The authors have analyzed the advantages of using geoinformation systems and spatial analysis tools for agricultural processes. The paper considers the effect of the soil hardness parameter on the yield of agricultural crops. Based on the data collected by the penetrometer, the authors formed the database. According to the results of the database analysis, using clustering methods, spatial models were created. The models display zones by average values of soil hardness. Based on the comparative analysis, the authors created a model for predicting maize yields based on soil hardness data. Based on this model, the authors selected optimal soil hardness values for each of the 60-centimeter intervals. For citation: Kolesnikov A.A., Yakovlev D.A., Poshivaylo Ya.G., Ivanov N.M., Kostić M., Bricheva A.V., Bambukh D.V. Maize yield forecasting with machine learning methods based on the results of soil hardness geoinformation analysis. Bulletin of the Tomsk Polytechnic University. Geo Assets Engineering, 2026, vol. 337, no. 5, pp. 21-35. https://doi.org/10.18799/24131830/2026/5/5368