Recurrent miscarriage (RM) and unexplained infertility (UI) are gordian knots in reproductive medicine, which are troubling many patients, doctors, and researchers. Although these two diseases of early pregnancy have a significant impact on human reproductive health, little is known about the specific mechanisms, which caused treatment difficulties. This study focused on the molecular signatures underlying the pathological phenotypes of two diseases, with the hope of using statistical methods to identify the significant core genes. An unbiased Weighted Correlation Network Analysis (WGCNA) algorithm was used for endometrial transcriptome data analysis and the disease-related gene modules were screened out. Through enrichment analysis of the candidate genes, we found similarities between both diseases and shared enrichment of immune-related pathways. Therefore, we used immune algorithms to assess the infiltration of immune cells and found abnormal increases of CD8+ T cells and neutrophils. In order to explore the molecular profile behind the immunophenotypic changes, we used the SVM algorithm and LASSO regression to identify the core genes with diagnostic capacity in both diseases and discussed their significance of immune disorders in the endometrium. In the end, the satisfactory diagnostic ability of these core genes was verified in the broader group. Our results demonstrated the presence of immune disorders in non-pregnancy tissues of RM and UI, and identified the core molecules of this phenotype, and discuss mechanisms. This provides exploratory evidence for the in-depth understanding of the mechanism of RM and UI and may provide potential targets for their future treatment.