Breast cancer (BC) is a leading cause of cancer-related mortality in females and is recognized as a molecularly heterogeneous disease. Previous studies have suggested that alternative messenger RNA (mRNA) processing, particularly alternative polyadenylation [poly(A)] (APA), can be a powerful molecular biomarker with prognostic potential. Therefore, in the present study, we profiled APA sites in the luminal B subtype of BC by sequencing APA sites (SAPAS) method, in order to assess the relation of these APA site-switching events to the recognized molecular subtypes of BC, and to discover novel candidate genes and pathways in BC. Through comprehensive analysis, the trend of APA site-switching events in the 3' untranslated regions (3'UTRs) in the luminal B subtype of BC were found to be the same as that in MCF7 cell lines. Among the genes involved in the events, a significantly greater number of genes was found with shortened 3'UTRs in the samples, which were samples of primary cancer with relatively low proliferation. These findings may provide novel information for the clinical diagnosis and prognosis on a molecular level. Several potential biomarkers with significantly differential tandem 3'UTRs and expression were found and validated. The related biological progresses and pathways involved were partly confirmed by other studies. In conclusion, this study provides new insight into the diagnosis and prognosis of BC from the APA site profile aspect.