The use of farm models to analyze agri-environmental policies requires selecting farms which can be hypothetical, typical or observed ones. Farm typologies, understood as a grouping of farms according to relevant farm characteristics, allow selecting most prevailing farm types for a modelling exercise. Thereby, a farm type represents a share of the real-word farm population. We develop a farm typology for the German Federal State of North Rhine-Westphalia based on the Farm Structure Survey 2016. It is designed to assess the revision of the German fertilization regulations in 2017 by applying a combination of a bio-physical crop model and a bio-economic farm model. The derived typology covers 77% of farms in North Rhine-Westphalia and comprises 210 farm types. Farms are grouped according to specialization, size in relation to area, and stocking density. In addition, a typical crop rotation is defined for every specialization in the nine soil-climate regions of North Rhine-Westphalia. We show that the proposed typology provides the necessary information for the selection of farm types as well as for the model initialization and parameterization in the described modelling exercise. Furthermore, we provide the information to adapt and extent the typology to similar research questions and upcoming Farm Structure Surveys. The incorporation of expert knowledge to identify farm structures which are not captured by the official statistic could improve the typology.