As with conventional farming, the improvement of organic farming systems requires agronomic planning tools to enhance economic performance. Crop rotation planning plays a crucial role in organic arable farming systems due to the renunciation of mineral nitrogen fertilisers and pesticides. Our objective was to develop a tool for generating and evaluating site-specific and agronomically sustainable crop rotations for organic farming systems in central Europe. The resulting static rule-based model, called ROTOR, consists of two basic steps: (A) A set of annual crop production activities (CPAs) is assembled semi-automatically from single site and crop-specific field operations using a relational data base. The database includes all relevant crops recorded separately with inputs and outputs, machinery and timing. Starting from stubble tillage and ending with the last harvest measure, the CPAs describe the current best cropping practices. Different CPAs are included for each crop according to (i) the type of crop preceding and (ii) the field operations following: whether ploughing or non-inverting tillage, undersowing crops, using catch crops, manuring, straw harvesting, or mechanical weed control. The former allows for the modelling of all possible positions of a crop within a crop rotation and the consequential effects of preceding crops. The CPAs are evaluated using rule-based assessment modules for yield, economic performance, N balance, nitrate leaching, and weed infestation risks. These modules have been developed using data from field experiments, farm trials and surveys, expert knowledge and a soil–crop simulation model. (B) Within the crop generation module, all possible sequences of CPAs are linked to 3–8-year preliminary crop rotations. Agronomically sustainable crop rotations are selected according to exclusion criteria (i.e., thresholds for N balance, weed infestation risks, phytosanitary and chronological restrictions) and ranked, e.g. by economic performance. The model was tested by comparing (i) estimated with observed yields and (ii) generated with existing rotations. These comparisons, based on data obtained from two farm surveys from North Eastern Germany, indicate the validity and usability of the model approach. ROTOR was found to support the complex crop rotation planning in organic farming systems requiring rotations with overlapping undersown main and cover crops. ROTOR is able to reduce the risk of planning failures by offering a quantitative method of optimisation of weed and site-specific N management.