The heterogeneity of pre-eclampsia (PE) complicates its pathogenesis, which remains incompletely understood. Emerging evidence indicates a significant role of metabolism in the pathophysiology of PE. We procured the PE dataset from the Gene Expression Omnibus database and sourced a published compilation of metabolism-related genes, then employed consensus clustering to classify PE subtypes. Subsequently, we examined the relationships of these subtypes with metabolic features and immune infiltration. Feature genes were identified using weighted gene co-expression network analysis (WGCNA) and further scrutinized through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses. To refine the selection of feature genes, we applied two machine learning algorithms. Additionally, we assessed the expression profiles of RAG1, RBBP7, RFTN2, SPATA7, and ZNF16 at the single-cell RNA sequencing (scRNA-seq) level. Finally, we validated the diagnostic value and expression of these genes using PE datasets and quantitative reverse transcription-PCR (qRT-PCR) analysis. We identified three PE subtypes on the basis of the number of distinct metabolic characteristics, namely Metabolism Correlated (MC) A (MCA), MCB, and MCC subclasses. Through WGCNA, we pinpointed 101 metabolic genes that were strongly associated with PE progression. Machine learning algorithms helped to narrow the list to five key signature genes, which were then used to construct a predictive model offering significant clinical benefits for PE patients. qRT-PCR analysis confirmed that these genes are closely linked to PE progression, while scRNA-seq data revealed high expression of RBBP7 in trophoblast cells. In conclusion, the five genes identified here-RAG1, RBBP7, RFTN2, SPATA7, and ZNF16-were found to be strongly associated with PE progression.