All over the world there is a strong interest and also potential for biogas production from organic residues as well as from different crops. However, to be commercially competitive with other types of fuels, efficiency improvements of the biogas production process are needed. In this paper, results of improvements studies done on a full scale co-digestion plant are presented. In the plant organic wastes from households and restaurants are mixed and digested with crops from pasture land. The areas for improvement of the plant addressed in this paper are treatment of the feed material to enhance the digestion rate, limitation of the ballast of organics in the water stream recirculated in the process, and use of the biogas plant residues at farms. Results from previous studies on pretreatment and membrane filtration of recirculated process water are combined for an estimation of the total improvement potential. Further, the possibility of using neural networks to predict biogas production using historical data from the full-scale biogas plant was investigated. Results from an investigation using the process residues as fertilizer are also presented. The results indicate a potential to increase the biogas yield from the process with up to over 30% with pre-treatment of the feed and including membrane filtration in the process. Neural networks have the potential to be used for prediction of biogas production. Further, it is shown that the residues from biogas production can be used as fertilizers but that the emission of N2O from the fertilized soil is dependent on the soil type and spreading technology. (c) 2012 Elsevier Ltd. All rights reserved.