Abstract: This article proposes a semantic-rank model for categorizing logical grammatical metaphors (LGMs) and analyzes their usage by Chinese EFL learners. The analysis is based on a corpus of English M.A. theses written by Chinese English learners and native English speakers. The findings reveal that learners tend to overuse nominalization and verbalization LGMs, exhibiting greater compression compared to native speakers. Additionally, learners demonstrate lower lexical variety in their use of LGMs. Issues such as the LGM syndrome, collocation, semantic prosody, and colligation are observed in the learners’ texts. The study provides pedagogical implications for advanced logical expressions.