Honeycomb structures are widely used in aerospace and other fields, which are prone to debonding defects during service. Therefore, it is necessary to detect the defects of honeycomb structures in time. In this paper, the square-wave thermography detection principle was studied. The thermal conduction of CFRP / Al honeycomb panel with debonding defects was simulated by the finite element analysis. Pulse phase thermography (PPT), principal component analysis (PCA), thermal signal reconstruction (TSR), and total harmonic distortion (THD) ware studied for the original infrared image sequence. The signal-to-noise ratio (SNR) was calculated and the processing results ware compared. PCA-PC1 had the maximum SNR, which could reduce the background noise interference in square-wave thermography. Square-wave thermography can accurately detect and identify the debonding defects of the honeycomb structure.