Abstract A fast and reproducible protocol for milk Nuclear Magnetic Resonance (NMR) metabolomic fingerprinting was developed, allowing for an accurate discrimination among milk samples from large-scale distribution, as well as among milk sample from different farms located in the same restricted geographical area. Seasonal variations in milk composition and correlations with cows' nutritional patterns are also assessed, underlining relationships between feeding and metabolites. The most important difference was related to the use of silage feeding. This finding is relevant to assess the suitability of milk for different dairy products. A prominent example is parmesan cheese, the preparation protocol of which excludes milk from silage-fed cows. Graphical abstract Unlabelled Image Highlights • NMR-based metabolomics allows for the characterization of biological matrices. • We developed a fast and reproducible protocol for metabolites extraction from milk. • We tested this protocol to monitor the origin of milk samples of different provenance. • Seasonal variations and influence of cow feedings were also explored. [ABSTRACT FROM AUTHOR]