Abstract: Due to the influence of interference noise, some minor defect features of the vehicle body weld seam may be masked, which affects the accuracy of defect detection. To address this issue, this paper proposes a non-destructive testing method for weld defects in railway electric locomotive body based on array eddy current technology. Using array eddy current technology, AC signals are applied in the welding seam area of the vehicle body to stimulate the generation of eddy current signals. By calculating the signal characteristics and expanding them into Fourier series. Combining normalization processing methods to highlight the characteristics of defect signals, thereby solving the problem of small defect signals being masked by noise. Using Maxwell's second equation system, obtain cross-sectional defect images of railway electric locomotive body welds. Analyze and detect defect size and depth through background subtraction processing. The experimental results show that the average defect size detection error of the detection results obtained by this method is 3.2 mm², and the average defect depth detection error is 1.7 mm, with high accuracy and excellent application performance.