In this paper, firstly, under the perspective of big data, the innovation research of piano intelligent teaching mode and the application strategy of intelligent piano teaching mode under big data thinking are explained, respectively. Secondly, in music recognition in a college piano classroom, the piano music signal needs to be sampled, quantized and pre-processed to avoid the frequency bias caused by the low-frequency components of the piano music signal. Then, based on the big data technology, the piano intelligent teaching note recognition system is constructed by using the improved time domain algorithm and the improved time-frequency domain combination algorithm, and the recognition of note time value in piano intelligent teaching and the piano intelligent teaching note recognition system are studied and analyzed. The results show that the improved time-frequency domain combination is more effective in responding to the time-value information of piano music signals in recognition simulation. The note recognition system studied in this project has seen a recognition rate increase of almost 5% in system performance. This paper studies the application of intelligent piano teaching in professional learning in universities, which can lead to the development and popularity of intelligent teaching methods in piano classrooms.