Biochemical Methane Potential (BMP) testing is an essential tool for deriving practical knowledge for optimizing and operating large-scale anaerobic digesters, monitoring, modeling and evaluating process performance, or when it is ongoing the development of a scenario analysis. While its usefulness is undoubted, the long duration of BMP testing is problematic for many of its applications, especially when timely results are required for decision making. In recent decades, numerous scientific contributions have demonstrated that a reduction in BMP test duration is possible by predicting the final gas production. The aim of this thesis work is the development of a new procedure/algorithm to obtain a preliminary estimate of the BMP experimental result using Machine Learning. The effectiveness and efficiency of the developed algorithm will be verified using experimental data from BMP tests performed on different substrates commonly fed to anaerobic digesters.