This study analyzes reviews of art exhibitions posted on SNS using text mining. The purpose of the study is to suggest implications for art production, exhibition planning, and art education. The target of the analysis is the exhibition appreciation text of Rosewylie’s and Yueminjun’s, which were posted on NAVER blogs among SNS. Among the posts posted during the exhibition period, only the appreciation written by visitiors was extracted and the first refinement was carried out. So 390 of the 2821 cases were collected in Rosewylie’s and 295 of the 916 cases were collected in Yueminjun’s. As a result, first, through the common keywords of the late two exhibitions, visitors focused on ‘color’ when viewing the exhibition, formed emotions, and subsequently engaged in thinking activities related to exploration. And they also focused primarily on searching for internal cues of the work. Second, the CONCOR analysis showed that visitor’s reviews formed clusters with the contents of the concept of art criticism, ‘description’, ‘analysis’, ‘interpretation’, and ‘evaluation’. Third, through N-gram analysis, visitor’s reviews formed a different form of text from the curatorial materials for the exhibition.