Under the environment of complex electromagnetic interference at the airport, high frequency weak signal produced by gas leakage of aircraft air-tightness system, can’t be accurately judged alone by testing instrument. For better developing useful signal discrimination, and improving testing effect of gas leakage signal, in the paper, wavelet threshold denoising algorithm and chaos system are applied. Through revising Donoho wavelet threshold denoising algorithm, defects of traditional denoising methods, like lower signal reconstruction precision, additive vibration and so on, are avoided. Signal to noise ratio of useful signal is further advanced. Through improving nonlinear recovery force item of Duffing oscillator equation, effective discrimination of high frequency weak signal is realized under the background of strong noise.