Volatile organic compounds (VOCs) constitute an important group of atmospheric pollutants emitted during various industrial processes, including energy conversion systems. Their presence in the atmosphere is associated with significant environmental and health impacts, as VOCs participate in photochemical reactions leading to the formation of tropospheric ozone and secondary organic aerosols, while many of them exhibit toxic or potentially carcinogenic properties. In thermal processes, VOC emissions are characterized by strong temporal variability and depend on numerous operational parameters such as temperature, fuel type, and process conditions. Consequently, conventional analytical approaches based solely on discrete measurements are often insufficient to fully capture the dynamic behaviour of emission processes. This study presents an approach combining analytical measurements with artificial intelligence methods for predictive analysis of VOC emissions in process gases. The research was based on a laboratory measurement system designed for the determination of VOC concentrations in gas streams generated during combustion processes. The experimental setup included a VOC detector, a sorption trap, and a controlled gas flow system, enabling measurements under laboratory conditions and continuous acquisition of emission data. The measurement system served as a source of experimental datasets used for further data-driven analysis. Based on the recorded process parameters and corresponding VOC concentrations, preliminary predictive models were developed using selected machine learning techniques, including artificial neural networks. The models allowed the identification of relationships between combustion process parameters and VOC emission levels, as well as the assessment of their predictive potential. The results demonstrate that the integration of laboratory measurement systems with machine learning methods provides a promising framework for the analysis and prediction of emission dynamics in thermal processes, supporting the development of advanced monitoring and decision-support tools in energy and environmental engineering.