In response to the comprehensive commercial fishing ban implemented under China’s “Ten-Year Fishing Moratorium” in the Yangtze River, which has led to challenges in distinguishing between wild and farmed aquatic products, this study utilized Hypophthalmichthys molitrix as a model species to systematically compare the fatty acid profiles in their muscle tissues and developed a discrimination model based on machine learning algorithms. First, by detecting and analyzing the characteristics of fatty acid composition in wild and farmed individuals, a discrimination system incorporating six machine learning algorithms was established. Subsequently, feature selection was applied to reduce the dimensionality of the original data, resulting in the identification of seven most discriminative feature fatty acids, which were used to construct an optimized model. The results demonstrated that dimensionality reduction significantly improved the discrimination performance across different algorithms, with AdaBoost.M1 exhibiting the best performance, achieving discrimination accuracies of 90.5% and 81.0% on the development set and test set, respectively. The findings indicate that fatty acid profiling combined with feature selection and machine learning algorithms enables high-accuracy discrimination between wild and farmed H. molitrix, providing a feasible technical approach for origin traceability of aquatic products. This approach offers profound support for the conservation of fishery resources and the enforcement of fishing policies in the Yangtze River Basin.