The feasibility of a combination method of surface-enhanced Raman spectroscopy (SERS) technology and linear regression algorithm was investigated for rapid quantitative analysis of pesticide residues in honey. The total of 30 samples was applied in the experiment with dimethoate pesticide residues range from 1 ppm to 10 ppm. The samples were divided into calibration set (20) and prediction set (10). The substrate of Klarite with an inverted pyramidal structure was adopted for improvement of the relative intensity of the majority of Raman shift peaks. The comparative analysis was carried out between SERS spectra of dimethoate pesticide residues in honey samples and conventional Raman spectra of dimethoate standard sample. And four characteristic Raman shift peaks at the wavenumbers of 867, 1 065, 1 317 and 1 453 cm-1 were found, which were related with the vibrational information of dimethoate molecule. The relationship was developed by linear regression algorithm between the intensity of Raman shift and the concentration of dimethoate pesticide residues. The 10 new samples in the prediction set were applied to evaluate the performance of the models. By comparison, the optimal model was obtained with the characteristic Raman shift peak of 867 cm-1. The higher correlation coefficient of prediction of 0.984 and lower root mean square error of prediction of 0.663 ppm were obtained. The detection limit of this method was 2 ppm, which was close to the maximum levels of pesticide residue detection limits. Experimental results showed that it was feasible to rapidly analyze quantitative of pesticide residues in honey with the combination method of SERS technology and linear regression algorithm. Compared with the conventional method coupled with the suitable pretreatment, the combination method of SERS technology and linear regression method could analyze the dimethoate pesticide residues in honey, and it also provided an optional method for rapid quantitative analysis pesticide residues in other agricultural products.