Purpose Despite the farm being considered by many as the most suitable level of decision-making and strategic management in agriculture, there is an increasing interest in evaluating agricultural management strategies at the regional level. Recent initiatives attempted to aggregate and generalise farm-level lifecycle inventory (LCI) data and lifecycle impact assessment (LCIA) results to describe the environmental performance of agricultural regions. This article describes our development and application of a regional statistics-based approach for constructing virtual representative farms (VRFs), representing dominant farm types for a given region, as a tool for comparing alternative regional agricultural strategies in contexts of insufficient farm (e.g. LCI) data. Methods Based on statistical sources, we constructed VRFs of the dominant farm types in the largely agricultural region of Brittany, France. Environmental impacts of different agricultural management strategies were estimated at the regional level by modelling the strategies as changes in VRF-based LCIs, calculating LCIAs and extrapolating their mean per-ha impacts to the total land use in the region. Based on this assessment, performed using a regional lifecycle assessment framework, we analysed relative environmental impacts of each management strategy on the region. A strategy-comparison table was built to allow decision makers to understand the potential regional environmental consequences of implementing each strategy. Results and discussion Once VRFs impact assessment results were extrapolated to the regional level, all strategies show environmental impacts per ha similar to those of the baseline, with differences ranging from -15 to +6%. The scenario featuring centralised fodder drying by 50% of cattle farms (50FOD) is the only one featuring surpluses for all products, due to associated cattle diet adjustments including reduced maize silage intake and partial substitution of concentrate feeds. The s