The article builds a music intelligent creation model by utilizing transform framework, and preprocesses music data through data embedding and location information encoding operations, deeply integrating technology and music creation. Combined with the encoder and decoder equipped with various attention mechanisms, the music intelligent creation is gradually realized through the steps of weight matrix calculation and linear transformation projection. In order to clarify the impact of the model on the future development of the music discipline, the model is applied in a music academy using classroom multimedia equipment as a medium. The results of the study are as follows: the music intelligent creation model enhances the feedback atmosphere of the traditional music classroom, and the average score of the students in each index is more than 3.9 points. The average scores of students’ aesthetic perception ability, artistic expression ability, cultural comprehension ability, and creative practice ability in the post-test were within the range of 3.88-4.52, which were 0.84-1.24 points higher than the scores in the pre-test. In addition, students’ average theory scores and music skill levels improved to 84.10 and 81.78 points, respectively. The results of the regression model showed that the classroom feedback atmosphere, music core literacy, and music theory and skills had highly significant positive effects on the future development of music discipline.
