"hanging a sheep's head and selling a dog's meat" often happens in real life. How to identify meat quickly and intelligently on site is a challenge. This research chooses a portable Raman spectrometer and a self-assembled film consisting of uniform gold nanoparticles as the signal amplification substrate. Using surface-enhanced Raman spectroscopy (SERS) technology, based on its high sensitivity, rapid, fingerprint identification and other features and advantages, combining with the intelligent algorithm of principal component analysis (PCA) and support vector machine (SVM). Achieving quickly and accurately identify of chicken, cattle, sheep and pig meat. This method provides an idea and a method for fast field testing of food safety.