<p>Generative artificial intelligence (GAI) has significantly transformed education, particularly by enhancing teaching capabilities. Central to maintaining educational quality is teachers’ disciplinary literacy, which underpins their ability to effectively integrate GAI tools into instruction. But few studies have rigorously examined pre-Service teachers’ disciplinary literacy in this domain. Addressing this gap requires examining the relationship between pre-service teachers’ disciplinary literacy and their use of prompt engineering. Within the discourse between humans and GAI, the BROKE framework, supported by prompt engineering, has gained extensive adoption. It comprises five fundamental components: Background, Role, Objective, Key Results, and Evolution. Through an eight-week quasi-experimental study, data were collected from 39 pre-service teachers, including questionnaire data on disciplinary literacy, dialogue data, and teaching design scores. The article first categorizes participants into low, medium, and high groups according to the intervals of 25%, 50%, and 25%, and describes statistical analysis from four dimensions: knowledge, application, evaluation, and ethics. Then, the article uses the BROKE framework to conduct epistemic network analysis(ENA) on the prompts of pre-service teachers at different levels. Finally, through correlation analysis and mediation effect analysis, it was found that subject literacy plays a partial mediating role in the improvement of pre-service teachers’ teaching design through prompt engineering, with a mediating effect of 31.36%. The research results indicate that appropriately using prompt engineering for human-computer collaborative dialogue can improve the subject literacy of pre-service teachers, further influencing the enhancement of teaching design.</p>