This paper is a corpus-based analysis of the semantic prosody and preference of “seem” to get a more fine-grained picture of the functions expressed by the marker. The multifunctionality of “seem” (appearance verb, evidential marker, hedging probability, indicating hearsay, etc.) has received a great deal of attention (Aijmer, 2009, Usonien? & Jolanta, 2013, Langacker, 2017, Marin, 2017, etc.) whilst few studies (if only) have looked on the issue of its profiled meaning thus this empirical research aims to explore the semantic prosody and preference of “seem” as an indicator to determine its primary function. The focus is on examining the semantic prosody, semantic preference and colligations of “seem” which help determines its predominant se-mantic function. In addition, how different meanings of “seem” interact with its various syntactic patterns and colligations is also investigated. It is argued that there is a close link between certain colligation and the preferred meaning it relates to.