Abstract: In this research, the surface enhanced Raman spectroscopy (SERS) technique is used to develop a fast detecting method for the detection of salbutamol in muscle tissues and liver. Silver colloids used for SERS are prepared by the reduction of silver nitrate with sodium citrate. 10 mL of hydroxylamine hydrochloride/sodium hydroxide solution (1.5×10-2 mol/L; 3×10-2 mol/L, respectively) to 90 mL of silver nitrate solution (1.11×10-2 mol/L), fully stirring and the silver colloids were obtained. Some constituents of muscle tissues and liver, such as protein, could significantly interfere with the SERS signal of salbutamol. Therefore, the ethyl acetate was used as the extraction solvent to precipitate the protein in muscle tissues and liver in alkaline (10% calcium carbonate solution), and the sample preparation method was simple and fast. This research was conducted based on a self-developed Raman system. The SERS spectra were obtained using 785 nm exciting radiation with 450 mW laser power and 5 s exposure time. Raman spectra of organic substances have severe fluorescence background, so it is crucial to remove fluorescence background from Raman signal first for subsequent signal analysis. In this study, Savitzky-Golay 5 points smoothing filter and the adaptive iteratively reweighted Penalized Least Squares (airPLS) correction method were used to remove the random noise and the fluorescence background for improving the accuracy of SERS results. In this research, we acquired the SERS spectra of salbutamol standard solution with different concentration from 10-3 mg/L to 5.5 mg/L. The SERS signals intensities decreased when the concentrations decreased. The peak intensity of salbutamol at 621, 814, 1 253, 1 489, 1 609 cm-1 could be used for monitoring the salbutamol levels. It could be found that these Raman peaks of salbutamol were still quite clear even at low concentration of 0.005 mg/L. Regression models showed a good linear relationship (R2=0.968) between the intensity of characteristic spectral peaks and concentration of salbutamol. The reproducibility of SERS detection was a very important parameter for SERS method. SERS signals of 50 muscle tissue samples with the salbutamol at the same concentration were measured to evaluate the reproducibility, and the relative standard deviation (RSD) of 50 parallel samples at 621, 814, 1 253, 1 489, 1 609 cm-1 were 6.54%, 6.07%, 8.65%, 7.44%, 6.81%, which indicanted good stability of the present method. Muscle tissues and liver samples were prepared with salbutamol at the concentration of 0.01-5 and 0.02-5 mg/kg, respectively. After the pretreatment of spectra, linear relationships were constructed to predict salbutamol concentration respectively. The lowest detectable levels for salbutamol and salbutamol concentration were 0.01 and 0.02 mg/kg for muscle tissues and liver samples, respectively. The recovery rate of muscle tissues and liver samples are 60.2%-75.3% and 64.4%-76.0%, respectively. Therefore, we constructed the models based on 5 characteristic spectral peaks intensity and salbutamol concentration in muscle tissues and liver samples, and the determination coefficients R2 were 0.912 and 0.921, respectively. The present study demonstrates a novel approach to detect salbutamol in muscle tissues and liver samples by using silver colloids, and the proposed method indicated a good potential for the evaluation of harmful additives in agricultural and animal product samples.