Purpose The purpose of this study was to evaluate the moderate survival data of porous tantalum rod implants for the treatment of osteonecrosis of the femoral head (ONFH). Additionally, some independent prognostic factors for conversion to total hip arthroplasty (THA) were identified. Methods The porous tantalum rod population was obtained from a prospective, consecutive group of patients treated for Steinberg stage I and II osteonecrosis from April 2009 through July 2011. The historical core decompression and impaction of bone filling particle subjects underwent surgery from April 2007 through March 2009. Surgical data including time of surgery, blood loss, and cell transfusions were recorded. Post-operative values were measured for hospitalization length as well as days requiring a patient-controlled analgesia (PCA) pump. Primary outcomes were Harris hip score and survivorship analysis. Demographics and baseline characteristics included age, sex, etiology, bilateral disease, associated chronic systemic disease, Steinberg stage, Harris hip score, accompanied with bone marrow edema of femoral head, and osteonecrotic lesion size. Results Demographic/baseline characteristics were similar between two groups. At the post-operative follow-up of 62 months, Harris hip scores were significantly increased (P < 0.0001) when compared to that before surgery in both groups. The magnitude of increase in the tantalum rod implant group was significantly greater than that in the control group (P = 0.0426). With an average follow-up of 48 months (range, 38-62 months), the tantalum rod group had an 84.6 % survival rate. With an average follow-up of 72 months (range, 67-85 months), the control group had a 63.3 % survival rate. A comparison of Kaplan-Meier curves showed significantly higher cumulative survival rates (P = 0.048) for hips with implantation of the porous tantalum rod (74.1 % at 62 months) than for those with impaction composite bone material (49.9 % at 62 months). The Cox proportional-hazard model revealed that implantation of tantalum rod (P = 0.012), bone marrow edema (P = 0.003), corticosteroids intake (P = 0.007), and age less than 50 years (P = 0.014) were the independent prognostic factors related to conversion into THA. Conclusions Compared with the traditional impaction composite bone material technique, implantation of tantalum rod in the treatment of Steinberg stages I and II ONFH can obtain better clinical results and higher cumulative survival rates. For patients without the use of corticosteroids, and especially for hips without bone marrow oedema, the clinical results from our study show highly encouraging survival rates and a delay in or prevention of conversion into THA.