Abstract Slurry constitutes an important substrate, increasingly often forming part of biogas production in biogas plants due to the significant content of methane in biogas produced from slurry. Slurry fermentation leads also to its deodorisation and significantly affects the sanitation process. Biogas production constitutes a microbiological process, one affected by many parameters, both physical and chemical. The complexity of the processes occurring during slurry fermentation means it is difficult to identify the significant parameters of a process. Therefore, the fermentation model is often defined as a “black box” method. Artificial neural networks (ANN) are becoming more frequently recognised as a tool to analyse processes that do not have a formal mathematical description (e.g. in the form of a structural model). Neural models enable one to conduct a comprehensive analysis of an issue, including in the context of forecasting biogas emissions during the slurry fermentation process. This study aims to develop a neural model that forecasts the level of methane emission during the slurry fermentation process. This study demonstrates that the generated neural predictor constitutes an efficient tool for estimating the amount of methane produced during bovine and porcine slurry fermentation processes.