In order to provide fresh concepts for the construction of a professional English teaching model, digital mining is used as a novel analysis tool. In order to teach professional English, this study first develops a combination of both online and offline education models based on the traits of English teaching. Second, using data mining the internet, we choose rough set theory as the primary algorithm for this study. We then analyze and decompose the information data table and build a rough set model for the mixed English teaching data. Finally, a process assessment and summative evaluation are used to analyze the hybrid online and offline English teaching software for professional English. The experimental class saw a 6.5% gain in scores, which was a much larger value and amount of improvement than those of the other seven classes. It shows that the learning effect of the online and offline blended learning style is better than the traditional learning mode.