When the Fast Fourier Transform (FFT) method is used to analyze the harmonics in a power network, the sampled signal’s non-integral period truncation and non-synchronous sampling can cause the spectrum leakage, which will affect the accuracy of the detective results. In order to improve the accuracy of the harmonic detection, this paper puts forward a Fast Fourier Transform algorithm based on four-spectrum-line interpolation algorithm. Through analyzing the adding window signal amplitude information at the discrete frequency points, the algorithm further enhances the accuracy of the amplitude and phase detection by using weighted arithmetic of four pillars spectral line near the harmonic frequency point. By data fitting, practical correction formulas corresponding to commonly used window functions are obtained. Through numerical simulation analysis, it can be seen that the algorithm has higher detection accuracy. The accuracy and effectiveness of the algorithm is verified by the results of the simulation.