Background: The burden of infection with antibiotic-resistant gram-positive cocci, including methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus (VRE), continues to increase, leading to substantial morbidity and high mortality rates, particularly in intensive care units (ICUs). Creative interventions may be required to reverse or stabilize this trend. Methods: The efficacy of empiric cycling of antibiotics active against gram-positive organisms was tested in a before-after intervention in a single surgical ICU. Four years of baseline data were compared with two years of data compiled after the implementation of a strategy where the empiric antibiotic of choice for the treatment of gram-positive infections (linezolid or vancomycin) was changed every three months. Whatever the initial choice of drug, if possible, the antibiotic was de-escalated after final culture results were obtained. The rates of all gram-positive infections were analyzed, with a particular focus on MRSA and VRE. Concurrently, similar outcomes were followed for patients treated on the same services but outside the ICU, where cycling was not practiced. Results: During the four years prior to cycling, 543 infections with gram-positive organisms were acquired in the ICU (45.3/1,000 patient-days), including 105 caused by MRSA (8.8/1,000 patient days) and 21 by VRE (1.8/1,000 patient-days). In the two years after implementation of cycling, 169 gram-positive infections were documented (28.1/1,000 patient-days; p < 0.0001 vs. non-cycling period), including 11 caused by MRSA (1.8/1,000 patient-days; p < 0.0001 vs. non-cycling period). The percentage of S. aureus infections caused by MRSA declined from 67% to 36%. The rate of infection with VRE was unchanged. Outside the ICU, the yearly numbers of infections with both MRSA and VRE increased over time. Conclusion: Quarterly cycling of linezolid and vancomycin in the ICU is a promising method to reduce infections with MRSA.