This paper first describes the main content and overall goal of the construction of the intelligent teaching platform under multi-dimensional data integration. Then, combined with the main content and overall goal, it implements the construction of the intelligent platform from three aspects: preprocessing of music information, characterization of music, and assessment of sight-singing ability, respectively. Finally, the correct matching rate of the selected chord sequence matching algorithm of the platform is verified through testing, the credibility of the platform’s sight-singing ability score is analyzed, and the students’ satisfaction with the platform is studied through a questionnaire survey. Compared with the other two algorithm models, the Chord Sequence Matching Algorithm improves the correct rate of recognizing chord types by a minimum of 2.35% and a maximum of 11.55%, which is a high correct rate. The experimental class consisting of 15 students verified that the scoring results of steps 1-4 of the platform scoring reached the expected goal of the test case with high credibility in the platform’s sight-singing ability scoring credibility test. The platform was used by the students participating in the experiment with a 100% satisfaction rate. The use of intelligent teaching platforms for music performance teaching can effectively improve the teaching effectiveness.