Abstract Background In septic arthritis, joint lavage can be performed using arthrocentesis (articular needle aspiration) or arthrotomy. The use of fluoroscopy to guide the puncture involves radiation. Ultrasound (US) guidance is still little recommended to guide the treatment of septic arthritis in children. We wanted to know whether treating septic arthritis in children was feasible and safe under ultrasound (US) guidance. Methods We retrospectively included 67 children (mean age, 3.0 years; range: 1 month–12 years) treated for septic arthritis of the hip, shoulder, or ankle using arthrocentesis or arthrotomy under US or fluoroscopic guidance (non-US group) with at least two years of follow-up. Results We found no significant difference between the groups. After arthrocentesis, patients in the US group remained in hospital for 0.8 days longer than those in the non-US group, but the difference was not significant. After arthrotomy, the arthrotomy-US group required 0.4 more days of hospitalization than the non-US group, but the difference was not significant. Patients in the US group exhibited higher initial CRP and WBC values than patients treated without US, although the differences were not significant. The WBC values of the arthrocentesis-US groups were higher than those of the non-US groups initially and at 72 h, but non significantly so; they became similar on day 5. Three puncture failures required arthrotomy (two under US guidance). Three patients required early revision surgery: one had undergone arthrocentesis with US, one arthrocentesis without US, and one arthrotomy without US. At the last follow-up, there were no clinical sequelae but two hip arthrotomies (one US and one non-US child) showed asymptomatic calcifications. Conclusions US guidance is feasible and safe for treating septic arthritis in children, visualizing structures not shown by X-rays and avoiding radiation exposure during surgery. Level of evidence IV (case series). Trial registration IRB-MTP_2021_05_202100781