In the recent years, the research on the dialogue system becomes more and more important. The performance of the automatic speech recognizer of the dialogue system is unsatisfied because of bad recognizing performance for spontaneous utterance. To solve this problem, this paper proposed an automatic speech recognition strategy directed by the semantic knowledge in dialogue system. The main idea of this strategy is to represent the expected utterance of speaker given by the dialogue management as a finite state network (FSN), which will be used for semantic parsing simultaneous with the speech recognizing and giving some feedbacks to direct the speech recognizing. The speech recognition strategy is tested on our dialogue system EasyFlight: a server of query and booking for the flight tickets. The experiment results show that the strategy has better performance than the baseline system.