There is no consensus on a diagnostic strategy for osteomyelitis underlying pressure ulcers. We conducted a prospective study to assess the accuracy of multiple bone biopsies and imaging to diagnose pelvic osteomyelitis. Patients with clinically suspected osteomyelitis beneath pelvic pressure ulcers were enrolled. Bone magnetic resonance imaging (MRI) and surgical bone biopsies (three or more for microbiology and one for histology per ulcer) were performed. Bacterial osteomyelitis diagnosis relied upon the association of positive histology and microbiology (at least one positive culture for non-commensal microorganisms or three or more for commensal microorganisms of the skin). From 2011 to 2014, 34 patients with 44 pressure ulcers were included. Bacterial osteomyelitis was diagnosed for 28 (82.3%) patients and 35 (79.5%) ulcers according to the composite criterion. Discrepancy was observed between histology and microbiology for 5 (11.4%) ulcers. Most common isolates were Staphylococcus aureus (77.1%), Peptostreptococcus (48.6%) and Bacteroides (40%), cultured in three or more samples in 42.9% of ulcers for S. aureus and ≥20% for anaerobes. Only 2.8% of ulcers had three or more positive specimens with coagulase-negative staphylococci, group B Streptococcus, and nil with enterococci and Pseudomonas aeruginosa. Staphylococcus aureus, Proteus and group milleri Streptococcus were recovered from one sample in 22.8%, 11.4% and 11.4% of ulcers, respectively. Agreement was poor between biopsies and MRI (κ 0.2). Sensitivity of MRI was 94.3% and specificity was 22.2%. The diagnosis of pelvic osteomyelitis relies on multiple surgical bone biopsies with microbiological and histological analyses. At least three bone samples allows the detection of pathogens and exclusion of contaminants. MRI is not routinely useful for diagnosis.