Australian grain-fed and grass-fed beef products garner premium market prices and many beef processors market branded beef supported by production system claims. Given beef cattle nutrition alters the fatty acid composition of subcutaneous fat and grass and grain-fed cattle can be differentiated using Raman spectra of the fat, this study aimed to evaluate if Raman spectroscopy and chemometric modelling can differentiate grass fed cattle from a variety of production systems. To this end, subcutaneous fat from a total of 520 beef carcases with 130 from long term grain-fed, short term grain-fed, grass-fed and supplemented grass-fed beef cattle were measured. Classification of carcases using Partial Least Squares Discriminant Analysis (PLS-DA) demonstrated spectra were able to correctly classify long term grain-fed (96%), short term grain-fed (85%), grass-fed (83%) and supplemented grass-fed (83%) carcases based on the test dataset. Spectral patterns including peaks that characterise fatty acids have been shown to underpin this classification. Overall, this study demonstrates Raman spectroscopy is a useful tool for the authentication and discrimination of beef carcases from different production systems.