To investigate the short- and mid-term effectiveness of revision hip arthroplasty by using impaction bone allograft and acetabular components in treatment of severe acetabular defects. A clinical data of 42 patients (44 hips) with severe acetabular defects between February 2011 and May 2018 were retrospectively analyzed. All patients underwent revision hip arthroplasty by using impaction bone allograft and acetabular components. Cemented cup (24 cases, 24 hips) and non-cemented cup (18 cases, 20 hips) were used in the revision surgery. There were 17 males and 25 females with an average age of 62.8 years (range, 22-84 years). The interval between the first total hip arthroplasty and revision was 2.5-12.0 years (mean, 8.3 years). The patients were accepted revision surgery for prosthesis aseptic loosening in 32 hips (31 cases) and the periprosthetic infection in 12 hips (11 cases). Twenty-nine hips (28 cases) were Paprosky type ⅢA and 15 hips (14 cases) were type ⅢB. The preoperative Harris score was 22.25±10.31 and the height of hip rotation center was (3.67±0.63) cm and the length difference of lower limbs was (3.41±0.64) cm. The operation time was 130-245 minutes (mean, 186 minutes) and the intraoperative blood loss was 600-2 400 mL (mean, 840 mL). The postoperative drainage volume was 250-1 450 mL (mean, 556 mL). Superficial infection of the incision occurred in 1 case, and the incisions healed by first intention in the other patients. All patients were followed up 6-87 months, with an average of 48.6 months. At last follow-up, the Harris score was 85.85±9.31, which was significantly different from the preoperative score ( t=18.563, P=0.000). Imaging examination revealed that the allogeneic bone gradually fused with the host bone, and no obvious bone resorption was observed. At last follow-up, the height of the hip rotation center was (1.01±0.21) cm, which was significantly different from the preoperative level ( t=17.549, P=0.000); the length difference of lower limbs was (0.62±0.51) cm, which was significantly different from the preoperative level ( t=14.211, P=0.000). The Harris score in the cemented group and non-cemented group increased significantly at last follow-up. The height of the hip rotation center decreased, and the hip rotation centers of both groups were within the Ranawat triangle zone. The length difference of the lower limbs also decreased, and the differences in all indexes were significant between pre- and post-operation ( P<0.05). There was significant difference in the height of the hip rotation center between groups ( t=2.095, P=0.042), but there was no significant difference in the Harris score and the length difference of lower limbs between groups ( P>0.05). For severe acetabular defect (Paprosky type Ⅲ), the hip can be reconstructed with the impaction bone allograft and cemented or non-cemented components in revision hip arthroplsty. The short- and mid-term effectiveness are satisfactory.