Protecting wheat from pests directly affects the country’s food security. Therefore, the purpose of this study was to create predictive models for estimating the harmfulness of E. integriceps by years. The harmfulness of E. integriceps was considered depending on the following indicators: pest abundance, environmental index (Wolf number) and hydrothermal moisture coefficient (HTC). The study proved the existence of mathematical uncertainty of information flows in relation to the specified pest, and therefore the mathematics of artificial neural networks with the structure of “multilayer perceptron” was used for forecasting. The results of the study of the harmfulness of E. integriceps to winter wheat in Ukraine were presented, including a forecast of the phytosanitary state of agrocenoses of Ukraine and recommendations for assessing the distribution of harmfulness of E. integriceps by years of observation (1996-2023) for the Odesa Oblast. It was noted that this distribution corresponds to a normal law with a mathematical expectation of 25%, which is confirmed by the results of observations for other regions of the Steppe zone. The relationship between the number of E. integriceps, Wolf number, and the accumulated integrated temperature and humidity characteristics of the environment was analysed. It was found that the harmfulness of E. integriceps is characterised by a fading periodic component with a period of 10-12 years. This result suggests the impact of the current year’s E. integriceps damage on the next year in 10-12 years. According to the forecasting results, the dependence of the harmfulness of E. integriceps on its number and the Wolf number was presented. Therewith, the accumulated integrated temperature and humidity characteristics of the environment were considered. The obtained findings are recommended for consideration in the organisation of planned technological operations for the protection of cereal grain crops