The purpose of this study is to review the transition and the trend of relevant research and the application of artificial intelligence on legal practice from 2000 to the end of March, 2019. This paper retrieved a total of 38 articles from Taiwan's master's thesis knowledge value-added system and the Taiwan journal article index system, and analyzed and compared the keywords, the years of publication and the legal practice fields they have applied. The current study found that most of the literature focuses on the application and discussion of the justice practice and patent. Moreover, we also found that “text mining” is the most widely adopted and frequently used related technology to AI, including Chinese segmentation and semantics, natural language, word frequency, etc. At the same time, the current study also found that the term “text mining” has gradually been covered by the concept of AI artificial intelligence or machine learning in the recent two years, which used to appear frequently in 2017 and 2018. Nowadays, more and more literature starts to mention “text mining” with artificial intelligence and machine learning in their studies, indicating that the relationship between these three will become more and more closely applied in legal practice. This phenomenon in the past helps us to understand the key points and the possible trend of AI applications on legal practice in the future.