For data mining courses in finance and economics universities, this study addresses the disconnection between professional teaching and value-oriented education by constructing a “knowledge point-ideological element” mapping and a "four-dimension" teaching scheme covering knowledge, competence, quality, and values. Based on teaching practices and data evidence from two classes, a blended teaching approach integrating case-driven learning and project-based evaluation is adopted to establish an implementation path of "knowledge mapping - case guidance - value enhancement". Using Accounting majors from classes 2023-3 and 2023-4 as samples, the study quantitatively compares overall scores, excellence rates, and standard deviations, supplemented by multidimensional indicators from teaching quality evaluations. The findings indicate significant improvements in excellence rates and overall satisfaction, with enhanced student awareness in data integrity, social responsibility, and technological ethics. However, issues such as class differences and insufficient natural integration of ideological elements are also observed. The study proposes constructing a case library based on financial contexts, adopting interactive classroom organization, and optimizing a "three-dimension evaluation" system (knowledge - competence - values), thereby providing a replicable path and evidence for integrating ideological education into data mining courses and improving teaching quality.