Information interaction and acquisition are now gradually breaking through the limit of time and space, Text detection technology based on semantic domain has become the new research in this field. This paper proposes topic link detection correlation model based on semantic technology, which involves the basic researches of text comprehension and semantic analysis in relevance determination. It is of great importance for subsequent researches in topic temporal detection. The core of this method is to establish the semantic structures for specific topics, by which the relevance determination can be achieved automatically reacting to the semantic information.