Abstract: Micro-leakage detection on gas pipelines has always been a tough problem in the field of leakage detection and fault diagnosis. In this paper, the leakage detection technology based on acoustic method was applied to detect microleakages of gas pipelines. First, the acquired signals were processed by means of wavelet transform to extract the acoustic characteristics which can propagate for a long distance. Then, the extracted acoustic characteristics were secondarily processed, and duffing chaotic oscillator was used to detect the existence of micro-leakage signals. Finally, the weak signals which propagated for a long distance were processed and verified. It is shown that the effective way to detect micro-leakages of gas pipelines is to process weak signals by means of wavelet-transform pre-processing combined with the chaotic oscillator. And it provides the basis for the application of acoustic leakage detection technology to gas pipelines.