Based on big data mining and learning analysis technology, this paper systematically summarizes the theoretical basis of network-based college students' English writing skills training mode and discusses the reform and application of college English writing teaching. By mining network data, establishing corresponding writing training database and writing corpus, and at the same time strengthening writing teaching methods with the help of existing network corpus. Students learn through reading, imitation and innovation in the course, and receive psychological and motivation training. The research shows how data mining ability can test the data production ability of front-line authors, and test the ability of front-line teachers to cultivate students' autonomous learning and self-adaptation. It is emphasized that this teaching mode updates the knowledge structure of front-line teachers and cultivates new teachers who combine knowledge with technology. By providing innovative methods to optimize the teaching process, it is expected to effectively improve students' writing ability.