This research is focused on studying the effectiveness of ML algorithms in predicting the probability of the Fusarium Head Blight disease occurrence in corn in Dnipro region of Ukraine. Linear regression, feedforward neural network and random forest models were considered for prediction. Current climate parameters (temperature, humidity, precipitation and leaf wetness duration) along with the climate parameters and disease occurrence probabilities for the last 3 h were taken into account for making predictions. The random forest model obtained the best metric score on the testing set: MAE = 0.47%, RMSE = 3.44% and R2 = 0.965.