Mastitis is an infectious inflammatory disease in dairy cows, causing economic losses and reducing animal welfare. With the final aim of discovering early and easily accessible biomarkers, the objective of this study was to validate the use of milk fat globules (MFGs) as a source of small non-coding RNAs (microRNAs, miRNAs) to diagnose mastitis. We studied the abundance of six miRNAs known to be regulated in the mammary gland during inflammation, by comparing their abundance in MFGs from five cows before and after lipopolysaccharide (LPS) challenge. Among them, four (miR-494-3p, -148a-3p, -99a-5p, -125b-5p) were differentially abundant according to inflammatory status. Bioinformatics analyses showed that they are predicted to target genes regulating cell life (e.g.: apoptotic process, cell cycle) and genes involved in gene expression (e.g.: mRNA processing, translation). Network analyses revealed that they are also related to Interleukin signaling network, confirming the contribution of miRNAs to the inflammation response of the mammary gland. We demonstrated that MFGs might be an easy source of miRNAs that are potential biomarkers to detect early mastitis allowing the application of a rapid and effective treatment. However, before its exploitation to diagnose mastitis, investigation of MFGs of early spontaneous mastitis must validate the use of these miRNAs as biomarkers of early infection of the mammary gland.