Food safety legislation plays a central role in regulating the levels of chemicals used in agriculture practices in order to prevent potential risks to consumers’ health within a certain region or country. Public Health organizations publish these regulations as recommendations on allowed quantities of chemicals residues for different types of crops. These documents pose a major challenge for automatic processing as their format is not normalized nor the terminology used is uniform in any way. Semantic Web technology tools offer a solution as these documents may be published as linked data which would allow computers to process them automatically, so that further analysis and interoperability would be possible. In this paper we introduce MRL-O, an ontology for describing data on allowed levels of residues present in commodities of agricultural origin. MRL-O serves as a standardized framework for sharing interoperable data and to provide tracking metadata about its sources and transformation processes. We also describe a step-by-step procedure to obtain MRL-O linked data from real non-normalized documents. Also, we applied this procedure on data published by Argentina and Brazil with promising results. Consequently, we argue that the proposed ontology is sufficient to model the domain of MRL regulation and serves as the basis for tools that support interoperability in this domain.