The dynamics of a construction project generates a huge amount of information for processing. Most project information is recorded in digital files and managed in a decentralized way. Current document management tools do not consider the semantic relationships between files and then cannot discover files based on their content or ensure information consistency across various information repositories. This paper presents a semantic-discovery method for accessing project information across repositories. By adopting semantic web technologies, including extensible markup language (XML), ontology, and logic rules, the semantic-discovery tool can access project files by their properties and relations. The method has been applied to searching for files using documents from a real project. The results show that the semantic content of a file can be used to ensure that only related files are retrieved. The dynamics of a construction project generates a huge amount of information for processing. Most project information is recorded in digital files and managed in a decentralized way. Current document management tools do not consider the semantic relationships between files and then can not discover files based on their content or ensure information consistency across various information repositories. This paper presents a semantic-discovery method for accessing project information across repositories. By adopting semantic web technologies, including extensible markup language (XML), ontology, and logic rules, the semantic-discovery tool The results has been applied to searching for files using documents from a real project. The results show that the semantic content of a file can be used to ensure that only related files are retrieved.