A power system harmonic detection technique based on a frequency domain interpolation wavelet transform is proposed. Aiming at harmonic detection, a fast Fourier transform (FFT) can only detect integer harmonics; a short time Fourier transform (STFT) can detect non-integer harmonics with low resolution; a wavelet transform can detect non-integer harmonics, but it has spectral aliasing and spectral leakage phenomena, which result in low resolution. This paper uses a frequency domain interpolation algorithm to detect a harmonic signal by choosing a Shannon wavelet. Simulation and comparative results obtained by three algorithms reveal that: the proposed method can effectively eliminate spectral aliasing, spectral leakage and the Gibbs phenomenon; it can precisely detect a harmonic signal, so it provides an effective means for power system harmonic analysis and detection.