Abstract: The acoustic pipeline leak detection system locates the leak position of pipeline by acoustic signal generated by the leakage. In order to detect pipeline leakage quickly and locate the leakage position accurately, it is important to detect the weak transient change of acoustic signal and the moment that the signal propagates to the acoustic sensor. Hilbert-Huang Transformation, as a time-frequency algorithm, has high time-frequency resolution. Based on this property, choosing Lanzhoou-Chengdu-Chongqing Products Pipeline as a simulation objective, Hilbert-Huang Transformation is used to process the leakage acoustic signals to get the time-frequency representation. By means of this method, frequency characteristic at each moment is analyzed and the accurate time that leakage acoustic signals get to the sensor can be obtained, and then the leakage location can be determined by calculation.