Abstract: This paper proposes a novel method of combination of conceptual graphs and simple semantic analysis, to solve the problem of low recall and precision rates.Using the formal of conceptual graphs to define the semantic orientation relationships of concepts in the text, and convert text filtering problem into conceptual graphs-based semantic matching problem.In order to simplify the process, reasoning matches with conceptual graphs on the basis of sensitive classification and words of part-of-speech.Practical application shows that the method can effectively identify and filter tendentiousness text message.