This article discusses the issue of diagnostics and predicting of faults in gas turbine units. The importance of the mentioned issue is due to the crucial role that gas turbine units play in various industries, and the need to ensure their reliable and cost-effective operation. The novelty of this research lies in the use of machine learning methods for diagnostics. This approach can outperform conventional statistical methods and, as shown in the article, allows for more efficient and accurate predictions. The investigation methodology included the creation and review of data sets on gas turbine unit health and defects using models. This approach allowed machine learning to be applied when actual data is scarce. The main objective of the research was to evaluate the effectiveness of using this approach in assessing the health of units, classifying a group of defects and analyzing the level of defect development based on gas thermodynamic parameters measured during operation. The results obtained show that the models can successfully classify fault groups and predict the stage of defect development. The achieved high accuracy in evaluating the health of gas turbine units confirms the potential of machine learning in improving diagnostics and maintenance systems. However, it is clear that there is still a need for further research using actual data to validate the findings.