This paper examines the phenomenon of conversational disfluency from the viewpoint of friendship and topics. Two research hypotheses were proposed, and various disfluency phenomena occurring in conversation were examined through a collected corpus conversation. The results did not reveal that conversations between people who were not friends or conversations that were not talked about necessarily led to disfluency. Thus it can be said that the phenomenon of disfluency does not arise from the tacit knowledge of some speech information. In this survey’s results, a ‘pause’ held a significant value in the group in which no friendship relationship had been formed. In addition, in the case of contents-centered words disfluency such as insertion or repetition, the number of occurrences was high in the conversation of the group that had assumed a specific topic. However, in general, the frequency of disfluency is higher than hypothesized when the level of friendship is high, or conversation is high in conversation group, and the occurrence of disfluency in conversation differs from that which occurs speaking alone . This paper discussed the phenomenon of disfluency focusing on small conversational corpus in the context of lack of discussion on the occurrence and type of non - In addition, although the occurrence of disfluency in a conversation is influenced by various contexts, this study faces certain limitations in that it examines the frequency and type of its occurrence mainly in the communicative situation and conversation of the participants.