This study investigates the effects of feedback modes on linguistic complexity of EFL undergraduates’ writing, with a focus on AWE (Automated Writing Evaluation) feedback and model text feedback. Using computerized tools such as LCA, L2SCA and Coh-Metrix 3.0,the study assesses ten commonly used measures in three lexical complexity dimensions and four syntactic complexity dimensions. Results reveal AWE feedback improves lexical complexity but decreases syntactic complexity, whereas model text feedback has positive effects on both lexical and syntactic complexity. The findings suggest that EFL students’ perception and adoption of feedback messages is a complicated cognitive process. Different feedback modes have intricate effects on linguistic complexity, and facilitate the improvement of text quality in different ways.