Aiming at the pain points of the construction of higher vocational English gold classes in the context of the integration of production and education, this paper explores the path of teaching quality improvement empowered by artificial intelligence technology. Students’ English pronunciation is captured using speech recognition technology, and learning style preferences are parsed based on SVM algorithm. The assessment method based on the standard speech reference template is proposed, and the pronunciation is corrected with the aid of the speech recognition results. The real-time pronunciation feedback and personalized resource recommendation mechanism are embedded into the smart teaching platform to promote the construction of English Golden Class. Setting up a 12- week teaching experiment, the mean difference between the scores of the two classes in the pre-test is 0.236, t=0.302, p=0.803>0.05, and there is no significant difference. In the posttest, the scores of the students in the experimental group were concentrated in the 35-50 point value range, while the scores of the control group were concentrated in the 20-35 point value range. The experiment proves that the proposed teaching model helps to promote the continuous deepening of the industry-teaching collaborative education model.
