In the field of periodic fatigue tests on metal structural components, existing data peak detection and correction methods suffer from low efficiency and poor adaptability in massive data processing. To address this issue, this study employs fiber Bragg grating strain sensors for the health monitoring of a metal structural component. Based on the data collected during fatigue tests, the problem of data errors caused by spectral distortion is resolved first. Subsequently, a peak data detection and correction method tailored for periodic fatigue test is proposed, which predicts peaks by utilizing the periodicity of data and achieves rapid detection and correction of peaks and valleys from the test data by combining frequency-domain analysis with statistical criteria. Comparative experimental results demonstrate that, compared with the sliding window extremum method, wavelet transform peak detection method, and K-nearest neighbors (KNN) density peak detection method, the method proposed in this paper exhibits higher detection accuracy and shorter processing time, enabling more effective data correction. This method provides strong support for efficient peak data detection and correction in the areas such as aircraft structural health monitoring and fatigue life evaluation.