Under the background of digital transformation, the integration and sharing of university English teaching resources face challenges such as low efficiency and insufficient matching. This paper takes data mining technology as the core and combines blockchain technology to construct an English teaching resource sharing system. A semantic feature segmentation method based on decision tree is proposed, and a support degree calculation model is designed to realize accurate recommendation and safe sharing of resources. Through experimental verification, the algorithm in this paper, compared with the α# algorithm with sub-optimal performance, improves the completeness by 6.93%, the accuracy by 20.07%, the symbolicity reaches 1.9017, and the performance is optimal in all three dimensions. After the system in this paper was put into use, students’ satisfaction with English teaching resources sharing increased from 54.56% to 81.62%. The one-way ANOVA of students in different majors has a significant difference, and the satisfaction of humanities majors is significantly higher than that of social sciences majors, P=0.017. The data-driven resource integration path can effectively solve the problem of teaching resource decentralization, and provide a technical reference for the digital transformation of education.
