Creating its own learner corpus, this research proposed to analyze and classify the transition markers used in solo presentation by 30 Thai engineering students based on Hyland (2019)'s Marker Categorization in Textual Metadiscourse. This research also aimed to identify and compare the quantities of individual transition markers among three groups of the selected 30 students (10 high, 10 mid, and 10 low users). The research employed both quantitative and qualitative method through SPSS, Transkriptor, and AntConc. The results showed that--and--was the top word utlized as a textual marker among all three groups while some other top transition markers ranked slightly differently among the groups. In addition to the search on the list of textual markers, the manual observation encountered other phrases acting as textual markers. The results indicated that users from average level can apply phrases as textual markers to communicate coherence instead of using traditional transition markers. Moreover, the "reminders", the act of reminding the listeners of previous content, manifested as a possible indicator of advanced proficiency users as the weak users were found to not produce any of this category. Thus, learners may improve their English-speaking skills if they are guided through the process of developing coherence in context using various categories of textual markers.