OBJECTIVE To evaluate the selection of the type of prosthesis in revision hip arthroplasty. METHODS There were 33 hips in our study, male in 7 hips and female in 26 hips. The average age of the patients were 59 years. The reasons of the revision included aseptic loosing in 22 hips, infection in 8 hips (2 infection hips with discharging sinuses), and acetabular erosion in 3 hips. The operations for revision were 13 cemented and 12 cementless acetabular prosthesis with autograft in morselized form; the femoral revision were all selected in cemented prosthesis. The revision for infection hip were all cemented prosthesis of extensively porouse-coated. RESULTS The average follow-up duartion was 3.9 years and 11 months. There was a radiolucency but no clinical instability accompanied in 2 hips and remaining moderate pain in 4 hips. No dislocation and fracture were seen in the series. Harris score were improved to 82.4 (68-88). CONCLUSIONS The commonest reason of revision hip arthroplasty was aseptic loosing. The acetabular prosthesis in revision could select cemented or cementless components and femoral prosthesis could select extensively coated stem. The cemented components could yield good results in infection hips revision.