Autochthonous ecotypes are plant genetic resources characterized by high genetic variability, tolerance to environmental stress conditions, the presence of specialized metabolites (terpenoids, flavonoids, and alkaloids), and health-promoting compounds. However, autochthonous ecotypes are gravely threatened by extinction, mainly because of their replacement by commercial varieties. To preserve autochthonous ecotypes and their diversity, as well as to make them an ideal alternative for commercial varieties, it is very important to identify their unique characteristics. The present study aimed to provide a global view of the metabolite diversity of three autochthonous lentil ecotypes from different villages of the Molise region (Italy): Capracotta, CA; Rionero Sannitico, RS; and Agnone, A; and one ecotype from the Umbria region (Italy): Castelluccio di Norcia, CS. Untargeted metabolomics, performed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS), allowed the detection of 663 differentially accumulated metabolites (DAMs), which were assigned to 13 metabolic categories using MS libraries (e.g., the GNPS Public Spectral Library) and a molecular network approach. Major classes included flavonoids (90–13.6%), amino acids and derivatives (70–10.5%), and cinnamic acids (55–8.3%). According to the PCA and PLS-DA score plots, the metabolite compositions of autochthonous ecotypes differed significantly from each other. These results have been confirmed by heatmap, allowing the selection of ecotype-specific metabolic features. The enrichment analysis is in progress to assess the presence of enriched metabolic categories for each ecotype, and the antibacterial, antiproliferative, and antioxidative activity of lentil extracts was also evaluated according to the differential metabolic profiles characterizing each ecotype.