Bacillus cereus is one of the pathogens causing nosocomial bloodstream infections (BSIs). However, few reports have documented the antimicrobial susceptibility and clinical characteristics of Bacillus cereus BSI and the importance of empirical therapy. The aim of this study was to investigate the clinical characteristics and antimicrobial susceptibility of B. cereus isolates from patients with BSI and to analyze the impact of appropriate empirical therapy on the outcome of patients with B. cereus BSI. All adult cases of bacteremia between April 2003 and March 2012 in a teaching hospital in Tokyo, Japan were reviewed retrospectively. Clinical data were collected from the patients' medical records and charts. Antimicrobial susceptibility testing was performed by broth microdilution method. The patients with B. cereus BSI were divided into an appropriate empirical therapy group and an inappropriate empirical therapy group. The primary outcome was all-cause mortality at 4 weeks after the start of BSI. The secondary outcome was early defervescence within 2 days after starting empirical therapy. There were 29 B. cereus bloodstream infection cases. No vancomycin, gentamicin, and imipenem-resistant isolates were found. However, 65.5 % were resistant to clindamycin and 10.3 % were resistant to levofloxacin. The main etiology was venous catheter-related (69 %). All-cause mortality at 4 weeks was not significantly different between the appropriate empirical therapy group (9 cases) and the inappropriate group (20 cases) in this study. However, early defervescence within 2 days after starting empirical therapy was significantly different (p = 0.032). The BSI of B. cereus is mostly caused by venous catheter-related infections. Appropriate empirical therapy is important to achieve early clinical resolution in B. cereus BSI. Vancomycin is one of the appropriate selections of empirical therapy for B. cereus BSI.