Objective To investigate the effects of antibiotic prophylaxis on the incidence of infection and prognosis in acute phase aftersevere stroke. Methods 165 patients within 24 h from clinic onset were divided into the intervention group and control group. The interventionincluded cefuroxime+metronidazole or moxifloxacin. The body temperature was continuously monitored, and the presence of infectionwas daily assessed with C-reactive protein (CRP), leukocyte count, and bedridden X-ray. They were also assessed with Glasgow ComaScale (GCS) and Acute Pathologic and Chronic Health Evaluation (APACHE)-Ⅱ. Primary end points was incidence of infection; secondaryend points included death and other clinical outcome. Results There was no difference between tow groups in infection rate (P=0.58), intervaluntil diagnosis of infection (P=0.74), so as mortality (P=1.00) and interval of death (P=0.84), body temperature (P=0.99), CRP (P=0.37) , leukocyte (P=0.51), scores of GCS (P=0.31) and APACHE-Ⅱ (P=0.28). Conclusion Antibiotic prophylaxis works little to prevent infectionand improve outcome in patients with acute stroke.