This paper is designed to improve the efficiency and quality of teaching and learning efforts and to meet the multidimensional information needs of music educators and learners. From the perspective of information fusion technology, we build a multi-source information fusion system based on the basic system architecture. The maximum likelihood, least squares, and D-S algorithms were used to calculate and construct the sampling evaluation index equations. A survey of the teaching methods used in music teaching in colleges and universities was conducted to analyze three strategies needed for college education reform: innovative teaching methods, strengthening teachers’ music competence and quality, and establishing feedback mechanisms for college music teaching evaluation. The combined survey results show that 72% of music teachers believe their teaching skills have been greatly enhanced, while students’ willingness to take music lessons after using information integration technology has all risen to over 80%.Therefore, in the process of music teaching in colleges and universities, the integration of information integration technology can not only improve students’ interest in learning but also improve teachers’ teaching ability, which provides a certain degree of reference for the reform of music teaching in colleges and universities.