Livestock production is one of the most water-use-intensive economic sectors globally, and pork is the biggest of all meat sectors, necessitating continuous improvement of the sector's water use. Environmental product declarations are one way of incentivizing environmental performance, but with the majority of the water use occurring in primary pig and feed production, methods are required that quantify the water use over the entire value chain. Life Cycle Assessment applies such an approach, and the European Commission's Product Environmental Footprint framework uses this methodology. Product Environmental Footprint studies can use generic data for the pig production and feed cultivation stages and results communicated without uncertainty. Current study aimed to test if using database water footprint inventories could lead to a systematic underestimation of the water use in Danish pork production. A probabilistic surface- and groundwater footprint inventory assessment of the production of 100 g pork in Denmark was carried out. Danish average industry data was used to assess the possible range of water use for domestic Danish processes and FAOSTAT- and Water Footprint Network data for imported feed. Monte Carlo simulations were used to create water footprint inventory intervals, which were compared with intervals for three inventory databases: EcoInvent, Agribalyse, and Agri-footprint. The water footprint inventory intervals for Danish pork ranged from 3.8 to 9.2 L/100 g with a coefficient of variation of 21%. Database values were significantly (p <0.001) left-shifted by 3.0–3.9 L/100 g and 4.4–6.6 L/100 g with significantly different (p <0.001) coefficients of variation of 6.4% and 12.3% for Agri-footprint and Agribalyse respectively. This makes using generic data preferable to using primary data for producers with low water efficiency. Instead of demanding primary data, it is recommended that uncertainties in databases capture the observable variability, and that environmental product declaration results must be communicated with their associated uncertainty. This could incentivize provision of primary data and avoid deliberate underestimations of water use.