Sound field analysis of oil well string is very important for the detection of oil well dynamic liquid level. Acoustic wave state is a key parameter, which indicates the relationship between oil well string and oil level. Because the echo resonance signal of acoustic wave contains the relevant parameters of oil level reflection, it needs to be processed. In this paper, the combination of principal component analysis (PCA) and Kalman filter algorithm is proposed to process the oil well echo resonance signal. Experimental results show that this method can significantly improve the noise filtering of the signal, and then improve the state estimation of the system more efficiently.