The ischemic cascade starts when atherosclerotic plaques decrease the supply of oxygen and substrates to cells and finalizes with myocardial infarction. These states have been here studied at metabolite level by optimization of a metabolomics profiling approach based on high-accuracy MS. For this purpose, serum samples from patients diagnosed with coronary artery disease and affected by stable angina or myocardial infarction (acute myocardial infarction or non-ST elevation myocardial infarction) were analyzed by LC-QTOF/MS after deproteinization to compare the profile of serum metabolites. The data set, composed by tentative molecular features detected in MS analyses, was filtered with statistical algorithms to remove entities resulting in redundant information. Tentative molecules were identified finding mainly lipids as statistically significant metabolites in the discrimination study due to their change in concentration. Lipids such as bile acid derivatives, phospholipids, and triglycerides were identified as relevant compounds for discrimination of individuals who suffered acute or non-ST elevation myocardial infarction from those suffering stable angina. The results achieved by this research could support the capability of metabolomics to go inside the study of artery diseases, and in addition to other omics disciplines could help to detect the occurrence of these disorders at initial stages or even to prognosticate their appearance.