Phage therapy has gained attention due to the spread of antibiotic-resistant bacteria and narrow pipeline of novel antibiotics. Phage cocktails are hypothesized to slow the overall development of resistance by challenging the bacteria with more than one phage. Here, we have used a combination of plate-, planktonic-, and biofilm-based screening assays to try to identify phage-antibiotic combinations that will eradicate preformed biofilms of Staphylococcus aureus strains that are otherwise difficult to kill. We have focused on methicillin-resistant S aureus (MRSA) strains and their daptomycin-nonsusceptible vancomycin-intermediate (DNS-VISA) derivatives to understand whether the phage-antibiotic interactions are altered by the changes associated with evolution from MRSA to DNS-VISA (which is known to occur in patients receiving antibiotic therapy). We evaluated the host range and cross-resistance patterns of five obligately lytic S. aureus myophages to select a three-phage cocktail. We screened these phages for their activity against 24-h bead biofilms and found that biofilms of two strains, D712 (DNS-VISA) and 8014 (MRSA), were the most resistant to killing by single phages. Specifically, even initial phage concentrations of 107 PFU per well could not prevent visible regrowth of bacteria from the treated biofilms. However, when we treated biofilms of the same two strains with phage-antibiotic combinations, we prevented bacterial regrowth when using up to 4 orders of magnitude less phage and antibiotic concentrations that were lower than our measured minimum biofilm inhibitory concentration. We did not see a consistent association between phage activity and the evolution of DNS-VISA genotypes in this small number of bacterial strains. IMPORTANCE The extracellular polymeric matrix of biofilms presents an impediment to antibiotic diffusion, facilitating the emergence of multidrug-resistant populations. While most phage cocktails are designed for the planktonic state of bacteria, it is important to take the biofilm mode of growth (the predominant mode of bacterial growth in nature) into consideration, as it is unclear how interactions between any specific phage and its bacterial hosts will depend on the physical properties of the growth environment. In addition, the extent of bacterial sensitivity to any given phage may vary from the planktonic to the biofilm state. Therefore, phage-containing treatments targeting biofilm infections such as catheters and prosthetic joint material may not be merely based on host range characteristics. Our results open avenues to new questions regarding phage-antibiotic treatment efficiency in the eradication of topologically structured biofilm settings and the extent of eradication efficacy relative to the single agents in biofilm populations.