Breast cancer (BC) is the most frequently diagnosed cancer and one of the major causes of female mortality. The development of prognostic models based on multiomics data is the main goal of precision oncology. Aberrant DNA methylation in BC is a diagnostic marker of carcinogenesis. Despite the existing factors of BC prognosis, introduction of methylation markers would make it possible to obtain more accurate prognostic scores. The study was aimed to assess DNA methylation signatures in various BC subtypes for clinical endpoints and patients' clinicopathological characteristics. The data on methylation of CpG dinucleotides (probes) and clinical characteristics of BC samples were obtained from The Cancer Genome Atlas Breast Cancer database. CpG dinucleotides associated with the selected endpoints were chosen by univariate Cox regression method. The LASSO method was used to search for stable probes, while further signature construction and testing of the clinical characteristics independence were performed using multivariate Cox regression. The dignostic and prognostic potential of the signatures was assessed using ROC analysis and Kaplan–Meier curves. It has been shown that the signatures of selected probes have a significant diagnostic (AUC 0.76–1) and prognostic (p < 0.05) potential. This approach has made it possible to identify 47 genes associated with good and poor prognosis, among these five genes have been described earlier. If the genome-wide DNA analysis results are available, the research approach applied can be used to study molecular pathogenesis of BC and other disorders.