Aims Conservative surgery (CS) for diabetic foot osteomyelitis (DFO) consists in removing all or part of the infected bone tissues without amputation, in complement with antibiotic therapy. Data on CS for DFO therapy are scarce. Material and Methods We performed a retrospective analysis of all DFO episodes treated with CS between 06/2007 and 12/2017. Remission was defined by the absence of soft-tissue infection, complete sustained (i.e. > 1 month) healing of the foot ulcer, favourable (i.e., stabilisation or improvement) radiological outcome, and no need for additional surgery during a 1-year follow-up. Results During the study period, 47 episodes (in 41 patients) were analysed. Excluding deaths (all unrelated to the DFO; n = 3) or loss to follow-up before 1 year (n = 5), the remission rate was 64.2%. Most failures occurred during the first 6 months (79%, 11/14). Patients who experienced failure had a higher rate of peripheral arterial disease with arterial stenosis than patients in remission (57% vs. 24%, P = 0.03), a higher C-reactive protein rate at admission (116 ± 112 mg/L vs. 48 ± 46 mg/L, P = 0.02), and a trend for a higher rate of abscesses (29% vs. 4%, P = 0.06). At 1-year follow-up, foot ulcers related to transfer lesion were identified in 25.5% of the cases. At the last follow-up (mean 3 ± 2 years), the remission rate was 23/25 (92%). Conclusions Our results suggest that CS is a therapeutic option in patients with localised but severe DFO. Clinicians should, however, consider the necessity of revascularisation, and higher risk of failure if surgery is performed in patients presenting with acute foot infections.