The objective of WP3 is to improve marine engine efficiency and reduce emissions by developing predictive models using chemical kinetic mechanisms and machine learning algorithms with a special focus on renewable fuels, i.e., methanol or ammonia. The special requirements for modelling marine engines, especially about the structure and parameterisation of the models, will be considered. The pollutant measurements and emission controls (WP1) and the dynamic measurements of in-cylinder pressure and gas temperature (WP2) will enable the model development by providing comprehensive measurement data. The developed models will be validated against the demonstration power unit(s) and refined by iterations. Furthermore, virtual sensor concepts will be developed to estimate difficult or even impossible quantities to measure or replace expensive measurement devices with low-cost sensors. Physical models always exhibit a certain degree of inaccuracy depending on the combustion system. The combination with machine learning models will serve to significantly reduce the degree of uncertainty. The achieved uncertainty reduction depends strongly on the quality and amount of the data. For each developed model in this WP, uncertainty quantification is done.