This paper explores using data mining in English teaching assessment in higher education within the 'Internet + Education' era. Traditional assessment methods struggle to meet modern teaching needs. By collecting diverse data like student performance and learning behavior, and employing data mining, a comprehensive assessment model is built. This model tracks student progress, uncovers data patterns, and offers personalized teaching guidance. Empirical analysis validates the model's feasibility, enabling colleges to better understand English teaching quality and adjust methods accordingly. In essence, this study offers a novel approach to assessment and contributes to education digitization and teaching enhancement.