An agri-food value chain (VC) represents a set of activities aimed at delivering highly valuable products to the market. Due to the diversity of actors in the agri-food VCs´ accumulated knowledge is typically situated within the boundaries of each entity of the VC. Hence, the question is how to improve knowledge sharing in agri-food VC, or more specifically how can knowledge flow and mobilize among different actors in the VC. To answer this question, we present a decision support system (DSS) for evaluation of knowledge sharing crossing boundaries in agri-food VC. The proposed DSS is developed through two phases: (i) identification of the most common knowledge boundaries by using machine learning and ontology technologies; (ii) transformation of the obtained ontology into a DSS for the evaluation of existing knowledge boundaries. In particular, the developed DSS helps in identifying, evaluating and providing directions for improvement of the knowledge sharing crossing boundaries in agri-food VC. We apply the DSS to evaluate three real VCs: a tomato VC in Argentina, a Chinese leaf VC in China and a brassica VC in the UK. The comparative analysis across the three varied case studies and their evaluation with the proposed DSS lead to more insights into knowledge-based decisions that a particular VC needs to address to improve its knowledge flow, in particular, to obtain insights in the transparency and interoperability of data and knowledge crossing boundaries in agri-food VCs.