High-yielding dairy cows face a stressed metabolic condition during the transition from late pregnancy to early lactation. To overcome this, complex adaptation processes are necessary in an orchestrated manner, affecting various physiological systems. To broaden our understanding of the biology underlying an efficient metabolic adaptation, this study aimed to characterize the changes of metabolic phenotype in healthy periparturient dairy cows in a targeted metabolomics approach. Twenty-six German Holstein cows were used to collect blood samples repeatedly during the transition period: 42 and 10 d before calving and 3, 21, and 100 d after calving. Blood serum samples were subjected to a targeted liquid chromatography-mass spectrometry (LC-MS) based metabolomics assay using the AbsoluteIDQ p180 Kit of Biocrates Life Science AG (Innsbruck, Austria). Processed metabolomics data were evaluated by principal component analysis (PCA) and by heatmap visualization to characterize metabolic patterns during the course of transition period. Further data analysis was applied to link these identified patterns with physiological and production performance data. A clear separation according to sampling days was revealed by the PCA, indicating a notable shift of the metabolic phenotype. According to the heatmap, acylcarnitine concentrations reflected patterns of high variation between individuals. Furthermore, the concentration of several glycerophospholipids and sphingolipids was remarkably decreased 10 d before and 3 d after calving than earlier and later in the transition period. Amino acids and biogenic amines showed a diffuse pattern with less variation over time. Metabolic pathways affecting specific lipid species, such as cell integrity, inflammation, and lipid trafficking, could be identified as targets for metabolic adaptation during transition from late pregnancy to early lactation. Analyzing longitudinal changes of the blood metabolome can help us improve our understanding of the multifaceted metabolic adaptation of transition cows, and can ultimately lead to defining and predicting healthy and unhealthy metabolic phenotypes in dairy cattle.