The progressive informatization of society allows us to receive an increasing amount of agriculture yield data and climate data. This allows to use of available data for more accurate forecasting of grain yields, grain prices, analysis of crop losses, etc. Climatic factors play a decisive role in grain yield fluctuations. Using climate databases makes it possible to build yield prognostic models, which allow in advance (in 3 months) to estimate the future yield. Pre-harvest yield forecasting can assist grain producers in making the necessary arrangements for the storage and marketing of the crop. In this work, a study of the influence of climatic factors on wheat yield fluctuations was carried out using machine learning techniques. We divided the detrended yield values into two groups, labeled as “low yield” and “high yield”. Five machine-learning models were trained on the available data and were used as classifiers. The random forest model and support vector machine are the most effective classifiers and provide 85% classification accuracy.