Objective: To explore the characteristic lipid signature in placentas collected from normal pregnancies and those with mild and severe intrahepatic cholestasis of pregnancy (ICP). This research aims to clarify the pathogenesis and identify lipid biomarker for ICP through LC-MS/MS based lipidomic analysis. Design: Cross-sectional study, including normal pregnancy women and women with mild and severe ICP. Setting: Chengdu Women’s and Children’s Center Hospital. Population: Placenta samples collected from 30 normal pregnancy women and 30 mild and severe ICP women respectively. Women with normal pregnancy and ICP were recruit from April 2021 to July 2022 in Chengdu, China. Main outcome measures: Differentially expressed lipids. Results: Fourty-four lipids were differentially expressed both in mild and severe ICP placenta. The pathway analysis revealed these lipids are mainly enriched in glycerophospholipid metabolism and autophagy pathway. Weighted correlation network analysis (WGCNA) identified the correlation network module of lipids highly related to ICP. Using multiple logistic regression analysis, we identified three and four combined metabolites that had an area under receiver operating characteristic curves (AUC) ≥ 0.90. Conclusion: Our results systematically revealed the lipid signature in mild and severe ICP placenta. The results may provide new insight into the treatment and early prediction of ICP.