Rifaximin for treating diarrhea-predominant irritable bowel syndrome (IBS-D) by regulating intestinal microbiota has been studied and recommended. In this study, we tried to investigate the effect of rifaximin on different components of intestinal microbiota and explore which component of gut microbiota can predict the efficacy of rifaximin in IBS-D. Healthy controls (HC) and IBS-D patients meeting the Rome III criteria were recruited, and IBS-D patients were orally administered 400 mg rifaximin three times daily for 2 weeks. Subjects were tested for small intestinal bacterial overgrowth (SIBO), their symptoms were recorded, and fecal and rectal mucosal samples were collected before and after treatment. Fecal and rectal mucosal bacterial data were obtained via 16S rRNA sequencing, and fecal fungal data were obtained via ITS2 sequencing. IBS-D patients were divided into two subgroups based on fecal bacterial composition, IBS1 (patients whose fecal bacterial composition were different from HC) and IBS0 (patients whose fecal bacterial profiles were similar to HC). Rifaximin increased fecal Bifidobacterium and decreased E. coli and Enterobacter in IBS1 patients. Although rectal mucosal bacteria and fecal fungi were not significantly altered in all patients after rifaximin intervention, rifaximin enhanced the connections among fecal bacteria, mucosal bacteria and fecal fungi in IBS1 patients. Compared with IBS0, we surprisingly found rifaximin ameliorated abdominal symptoms of IBS1 much better. Receiver operating curve analysis revealed patients whose fecal microbial dysbiosis indices (MDI) were higher than -3.006 could be diagnosed as IBS1. Fecal bacterial dysbiosis could be a biomarker for rifaximin treatment for IBS-D.