Fish oil is one of the health products pursued by the public at present. The volume of the domestic fish oil consumption market is also growing, and the adulteration of fish oil is also becoming increasingly prominent. To realize the quantitative analysis of fish oil adulteration, this paper proposes a research work based on near-infrared micro-Raman spectroscopy combined with an elastic net quantitative inversion algorithm to realize the adulteration of fish oil with different animal oils. The Raman spectrum characteristic peaks of different animal oils were analyzed in detail. The Raman spectrum databases of adulterated fish oils and the corresponding ElasticNet quantitative inversion models were established. The results showed that the R2 of the four quantitative models for the test set were 0.9848, 0.9876, 0.9886, and 0.9880, and the RMSE was 0.0389, 0.0352, 0.0339, and 0.0347, respectively. Therefore, for the problem of fish oil adulteration, the technical method of combining ElasticNet quantitative inversion algorithm and Raman spectroscopy proposed in this paper has essential reference value and research significance for the field of fish oil quality detection, and this method can guide the application of in-situ rapid detection of fish oil detection in the future.