One-dimensional power quality signal usually contains different degrees of white noise, and the efficient denoising is an important prerequisite for the detection and identification of power quality signals. In order to effectively eliminate the noise and restore the singular points of the signal, a new denoising method based on Complete Ensemble Empirical Mode Decomposition with adaptive noise (CEEMDAN) and wavelet adaptive threshold is adopted. Firstly, the self-correlation method is used to sieve the high-frequency Intrinsic Mode Functions (IMFs) with noise decomposed by CEEMDAN. Then, wavelet adaptive threshold denoising is performed on the high frequency components, which preserves the effective information of the high frequencies. Finally, signal reconstruction is performed with other IMF components. The simulation results show that the proposed method has a good denoising effect and can effectively keep the real information in the high frequency components, which provides a new idea for the effective denoising of the power quality signal.