To seek the path and method of innovation and reform of art practice of ethnic music education in colleges and universities. Based on information fusion technology, this paper introduces the Kalman filter algorithm for analysis and recognition of ethnic music education in colleges and universities and understands the application of information fusion technology in ethnic music education in colleges and universities through the correct recognition rate and recognition time of music signals. We also apply independent experiential and problem-based experiments to verify the feasibility of information fusion technology in innovative reform. The experimental results show that the Kalman filter algorithm based on information fusion technology has a 96.57% correct recognition rate for ethnic music signals, and the recognition time is faster than the support vector machine and BP neural network by 3.587s and 1.291s on average. 73.47% of the students, on average, think that the innovative reform of art practice based on information fusion technology in college ethnic music education is very effective, which is higher than the original The average percentage of students who thought the effect of the innovation reform of folk music education based on information integration technology was very good was 73.47%, which was 60.37 percentage points higher than the original research results. The average percentage of students who thought it was ineffective was only 9.19%, 52.31 percentage points lower than the original research results. The above results prove the feasibility of information fusion technology in reforming art practice innovation in university ethnomusicology education and provide a new direction for reforming art practice innovation in university ethnomusicology education.