Numerous studies currently compare the lipid metabolism in patients with cardiovascular disease (CVD) and healthy individuals to identify lipid markers for predicting CVD. In this study, multidimensional mass spectrometry-based shotgun lipidomics was used to examine the serum lipidomics of participants in a clinical randomized controlled feeding trial undergoing olive oil (OO), camellia seed oil (CSO), and soybean oil (SO) dietary interventions. 189 lipid molecules are identified, including 14 species of phosphatidylinositol, 45 species of ethanolamine glycerols (PE), 47 species of choline glycerophospholipids (PC), 39 species of triacylglycerols (TAG), 18 species of lysophosphatidylcholine, and 26 species of sphingomyelin. After screening, 10 lipid markers are found, among which 18:2 fatty acid (FA), 16:1 FA, C54:4/C55:11, C54:3/C55:10, and C52:3/C53:10 in TAG pool, p18:0/20:0 and a18:0/18:1 in PC pool, and p18:1/20:4 in PE pool have differential regulation in the SO group compared to OO and CSO. The d16:0/18:1 in PC pool and C52:2/C53:9 in TAG pool are differentially regulated by OO and CSO. The C52:2/C53:9 in TAG pool has a significant negative correlation with aspartate aminotransferase (r = -0.363, P = 0.048) and high-density lipoprotein cholesterol (r = -0.519, P < 0.01). This study provides a reference for researching the effect of dietary fat on blood lipid metabolism.