Automated writing evaluation (AWE) systems have been introduced to ESL/EFL classes in the hopes of reducing teachers' workloads and improving students' writing by providing instant holistic scores and corrective feedback (Jiang & Yu, 2020; Link et al., 2014; Ranalli & Yamashita, 2019; Warschauer & Ware, 2006). When it comes to genre-specific writing, general AWE feedback may be insufficient because communicative purposes should be achieved, for which feedback is needed beyond grammar and mechanics. However, very few genre-specific AWE systems based on rhetorical move analysis have been developed. Therefore, the present study reports on the development and evaluation of a genre-based AWE system to facilitate Taiwanese engineering graduate students' writing of research abstracts. This AWE system provides automated feedback on two linguistic features, lexical bundles and grammatical categories of verbs (i.e., tense, aspect, and voice), associated with moves in abstracts. The feedback was designed to be co-constructed between learners and computers in order to promote interaction. The effectiveness of the AWE system was evaluated following Chapelle's (2001) computer-assisted language learning evaluation framework. The findings revealed positive effects; with appropriate guidance, the AWE system was able to draw participants' attention to and enhanced their use of these two linguistic features to achieve the communicative purposes of rhetorical moves in their abstracts.