Abstract: Power communication optical cables are key infrastructure of the smart grid, and their operational status directly affects grid security. To address the issues that power optical cable lines are prone to anomalies such as bending and fiber breakage in complex electromagnetic environments, and that traditional OTDR detection methods suffer from severe noise interference and low fault localization accuracy, a wavelet transform based anomaly detection method for power optical cable lines is proposed. First, wavelet transform is used to perform multi-scale decomposition of the OTDR backscattered signal, and an improved soft threshold function is adopted to remove high-frequency noise. Second, the wavelet transform modulus maxima (WTMM) is employed to detect signal singularities, and anomaly localization and type discrimination are achieved by combining typical fault characteristics. A 10 km single-mode fiber simulation model is built on the MATLAB platform, where two typical faults (bending loss and fiber breakage) are set, and Gaussian white noise with a signal-to-noise ratio (SNR) of 15 dB is added for validation. The results show that the proposed method can improve the SNR to 24.2 dB, achieve an average localization error of 10.5 m for the bending point and fiber break point, and attain an average detection accuracy of 95%, outperforming traditional moving average filtering and direct threshold methods. This method provides effective technical support for online intelligent monitoring of power optical cables.