Surface-enhanced Raman spectroscopy (SERS) of chicken was collected by DXRTM micro-Raman spectrometer with gold colloid as an active substrate and NaCl solution as the active agent. Rapid identification of sulfadimidine (SM-2) and sulfadiazine (SPD) residues in chicken was achieved. Raman peaks at 937 and 1 188 cm-1 were used to determine whether there are SM-2 and SPD in chicken or not. The Single-factor experiment method was used to optimize the experimental conditions, and the optimum experimental conditions were obtained: the addition amount of Gold glue was 500 L, the addition amount of NaCl solutionwas 100 L and the adsorption time was 5 minutes. The original Raman spectra were pre-processed by adaptive iterative penalty least squares (air-PLS), normalization and second derivative. Then the eigenvectors were extracted by principal component analysis (PCA). Finally, the first four PCA scores were used as input values of the support vector machine (SVM) classification model, and the SVM classification model based on C-SVC type was established. The optimal penalty parameter c was 0.01, and the kernel parameter g was 0.1. The overall classification accuracy of the model was 93.23%, the sensitivity of chicken containing SM-2+SPD was 100%, and the specificity of chicken containing SPD was 99.02%. The results showed that this method had good identification effects. It could be used to detect and identify SM-2 and SPD antibiotic residues in chicken quickly.