该文给出了连续频谱和离散频谱的谱质心计算公式，讨论了谱质心应用于频谱校正的原理和步骤。通过计算离散频谱正频率段的谱质心估计信号频率，根据邻近谱质心的峰值谱线估计信号幅值和相位。对于谐波信号，引入子带谱质心概念，根据谐波次数将谐波分成多个子带，计算每个子带的谱质心估计各次谐波参数。仿真实例表明，谱质心法能提高谐波分析的准确度，有效克服基波频率波动和噪声对谐波分析的影响。与传统的比值法、相位差法和能量重心法相比，谱质心法可以避免繁琐的峰值搜索和复杂的谱线插值计算，不需要推导窗函数的能量特性，是一种简单实用的电力谐波分析方法。“,”Spectral centroid (SC) formulas were given for both continuous spectrum and discrete spectrums. SC’s applications theory and steps in spectrum correction were discussed. The signal frequency was estimated through computing SC within positive frequencies of the discrete spectrum. The signal magnitude and phase were estimated by the peak spectrum line near SC. For harmonic signals, the sub-band spectral centroid (SSC) was introduced. The harmonic signal was divided into a number of sub-bands according to harmonic times. Each harmonic parameter was estimated after computing SSC. The simulation examples show that the proposed method can improve harmonic analysis accuracy and can overcome effects of the fundamental frequency fluctuation and noise. Compared with the tradition ratio method, the phase difference method and the energy centroid method, the proposed method can avoid the troublesome peak search and the complex spectrum line interpolation, and need not derive energy features of window functions, which is a simple and practical method for power harmonic analysis.