<p>A relevant direction in modern medicine is the development of non-pharmacological methods of pain relief. The use of information systems and resources, such as virtual reality technologies, for implementing such methods to reduce pain sensations has high potential. On the other hand, the effectiveness of using virtual reality must be objectively substantiated and proven. To investigate the impact of various distracting factors, including virtual reality technologies, on the level of pain sensations. A method for assessing the intensity of pain sensations was proposed, based on the combination of a set of characteristics from electroencephalograms (EEGs) of brain activity and survey results from participants using an 11-point numerical pain rating scale. Seven tests were formulated, including a resting state, exposure to a pain source, and the addition of various factors (music, video, virtual reality). For each test, analysis and processing of the obtained EEG data were suggested to calculate the spectral power density, coherence matrices, Fractal Dimension analysis, and the degree of interhemispheric asymmetry. The task of classifying human pain sensations was considered using 16 different machine learning algorithms on 10 different sets of initial data (differing in information volume and the number of analyzed EEG characteristics). Data collection and analysis of EEGs across 7 tests were conducted with a group of 23 individuals (21.5 ± 4.8&#xa0;years). Selected metrics were calculated, and comparisons were made among the 10 data variants for the 16 machine learning algorithms. As a result of the comparison, the best performance across all experiments was achieved using the Stacking Classifier algorithm, which combined several simpler classifiers (Logistic Regression, KNeighbors Classifier, Decision Tree Classifier, MLP Classifier, and Quadratic Discriminant Analysis). In the experiment determining the presence or absence of pain impact, this algorithm demonstrated an accuracy of 85%, while classifying all 7 types of tests yielded an accuracy of 72.5%. The conducted research partially confirmed the hypothesis regarding the possibility of using EEG data to assess human pain sensations and their intensity. It fully confirmed hypotheses related to evaluating the influence of various distracting factors on pain levels and the applicability of different data variants for solving the classification problem of human states. The obtained results may find application in clinical practice as one of the tools for reducing pain sensations.</p>