Piano timbre recognition and intelligent synthesis are of great significance in realizing the intelligent teaching of piano timbre. This paper takes the piano timbre teaching based on artificial intelligence interaction as the research object, constructs the timbre expression spectrum based on harmonic structure through the exploration of timbre synthesis, timbre features and other related theories, proposes the timbre feature extraction method based on the time-frequency cepstrum domain of the piano music signal, and then constructs the piano timbre recognition and intelligent synthesis system, realizes the simulation of the piano music, and then provides an intelligent interactive tool for the piano timbre teaching. The method is used to construct a piano tone recognition and intelligent synthesis system. When using the method in this paper, the amplitude of the piano tends to be stable when the frequency is 1600Hz~2400Hz, and there is no noise interference, and when the frequency is 2500Hz and 2800Hz, the amplitude is the lowest, and the recognition performance of the piano timbre is better. Meanwhile, the correct rate of timbre recognition of this method reaches 87.83%, which is better than 58.54% of the comparison method. In addition, the musical tone signals simulated by the method in this paper are very close to the theoretical values of each note of the real piano instrument captured, with an accuracy rate of up to 99%, which proves the accuracy of the simulated piano sounding. And the method can effectively promote the combination of artificial intelligence technology and piano teaching concept, the confidence level of quantitative regression analysis is high, and the evaluation results of teaching quality are good, which provides a reliable theoretical and practical basis for realizing the high-quality teaching of piano timbre.