Abstract: Wingbeat frequency is one of the important parameters of insect radar for insect target recognition. Because of the weakness of insect wingbeat signal, in high background noise, the echo has low signal-to-noise ratio (SNR) and weak signal strength, which affects the accuracy of insect wing vibration frequency extraction. In this paper, we proposed an extraction method of insect wingbeat frequency under low SNR. The method used the multi-resolution analysis characteristics of wavelet transform to analyze the relationship between the echo wavelet coefficients and the noise wavelet coefficients at each scale, determined the wavelet decomposition level, denoised the signal by wavelet threshold. Then the insect wingbeat frequency was extracted by using the time-frequency analysis, the accuracy of which is improved. The results of simulation analysis and experimental data were processed to verify the validity of the proposed method.