Abstract: In order to extract the main characteristic frequencies of the acoustic monitoring signal of the leakage of heating pipelines more accurately, a signal denoising method is necessary. In this paper, variational mode decomposition (VMD) was employed to decompose the original acoustic signals collected at spots 5 cm and 40 cm away from the leakage spot, respectively. The permutation entropy of each modal component was calculated and was used as the standard for removing the noise signals. Finally, the signal was reconstructed. The processing results were compared with the results processed by the empirical mode decomposition (EMD) and the ensemble empirical mode decomposition (EEMD). Compared with the other two decomposition noise reduction methods, VMD can solve the modal aliasing problem better and remove the noise signal more accurately.