Lexical bundles have attracted significant attention from ESL researchers, concerning their role in improving student writing performance, so various quantitative and qualitative analysis have emerged in this field. Limited contrastive studies investigate the use of LBs in written essays at tertiary levels; hence, this corpus-based study is conducted to identify the most commonly used LBs in English texts written by L1-Vietnamese and L1-English undergraduates, to examine how these LBs differ in structural and functional uses, and to compare how much these LBs align with those listed in the AFL (Simpson- Vlach & Ellis, 2010). To achieve these, firstly, a dataset of two learner corpora, namely an unpublished L2-Vietnamese corpus (55,671 words) and a (sub)-corpus from LOCNESS (Granger, 1998) used as L1-English corpus (52,980 words) was analyzed. Secondly, given that LBs, according to Biber et al. (1999), represent key constitutive elements that influence user writing proficiency, the analysis of structures and functions of the most frequently used LBs was carried out. Thirdly, the AFL is used to compare with two frequency lists. The results indicate that L1-Vietnamese students deploy more LBs than L1-English ones, with total normalized numbers per 10,000 words of 40 vs. 28, and frequencies of 395 vs. 243, respectively. Both groups of students use ‘Noun phrase with of-phrase fragments’, ‘Other prepositional phrases’, and ‘Pronoun/ Noun phrase + be/verb phrases’ as the top three frequent patterns. Regarding discourse functions, both show a preference for using referential bundles. While L1-Vietnamese undergraduates often convey abstract concepts and quantitative attributes, L1-English students prefer using bundles to emphasize and express main points. Moreover, there is some degree of overlapping LBs, indicating that the AFL may include commonly used LBs being relevant in both contexts, regardless of the domains or English proficiency. Consequently