This paper extracts the audio signals during piano playing, and uses two main audio signal processing techniques, namely audio signal recognition and wavelet transform noise reduction, to pre-process and extract features of the piano audio, and analyzes the influence of the audio signal processing techniques in the expression of music aesthetics by combining with the simulation experiments. The wavelet analysis method used in this paper has a signal-to-noise ratio of 7.55 at decomposition layer 7, and the relative error is 0.16. The model in this paper can predict the playing instruments according to the timbre features, and the accuracy of this model for the recognition of Happy, Angry, Sad, Fear, and Neutral emotional expressions is 0.001. The model in this paper can predict the musical instruments according to the timbre characteristics, and the recognition accuracy of the model for Happy, Angry, Sad, Fear and Neutral is between 0.91 and 0.947. The sound quality of the piano performance before and after the audio signal processing technology has greatly improved, with the SNR increasing by 13dB and the THD decreasing by more than 60%. Audio signal processing technology has the potential to enhance the accuracy of audio signal recognition in piano performance, as well as enhance the sound quality effect and enhance the expression of music aesthetics.