Introduction Although it is still at a very early stage compared to its mass spectrometry (MS) counterpart, proton nuclear magnetic resonance (NMR) lipidomics is worth being investigated as an original and complementary solution for lipidomics. Dedicated sample preparation protocols and adapted data acquisition methods have to be developed to set up an NMR lipidomics workflow; in particular, the considerable overlap observed for lipid signals on 1D spectra may hamper its applicability. Objectives The study describes the development of a complete proton NMR lipidomics workflow for application to serum fingerprinting. It includes the assessment of fast 2D NMR strategies, which, besides reducing signal overlap by spreading the signals along a second dimension, offer compatibility with the high-throughput requirements of food quality characterization. Method The robustness of the developed sample preparation protocol is assessed in terms of repeatability and ability to provide informative fingerprints; further, different NMR acquisition schemes-including classical 1D, fast 2D based on nonuniform sampling or ultrafast schemes-are evaluated and compared. Finally, as a proof of concept, the developed workflow is applied to characterize lipid profiles disruption in serum from beta-agonists diet fed pigs. Results Our results show the ability of the workflow to discriminate efficiently sample groups based on their lipidic profile, while using fast 2D NMR methods in an automated acquisition framework. Conclusion This work demonstrates the potential of fast multidimensional H-1 NMR-suited with an appropriate sample preparation-for lipidomics fingerprinting as well as its applicability to address chemical food safety issues.