Background: The modular hemipelvic prosthesis has been used in patient of Type I-IV pelvic tumor with good outcomes, but how to keep the stability between the prosthesis and the residual sacrum is a problem. An additional screw-rod system seems to solve it, but its biomechanical characters are still not well understood, which need experimental evaluation.Methods: Six pelvic specimens were prepared in three conditions(normal intact pelvis, “normal”; the pelvis of left Type I-IV defect and implanted with prosthesis without/with additional screw-rod system, “rod-” and “rod+”). Compressing biomechanical experiments(50-500N) were performed in these three conditions, respectively.Results: The loadings during the experiments are in accordance with the linear elastic control mode. Under the increasing loading, the implanted pelvises displaced asymmetrically, unlike normal intact pelvis. The vertical displacement of “rod+” changed significantly, whereas “rod-” did not. For both implanted pelvis, right side displaced less than left side(P values <0.05).Conclusions: The implanted pelvis showed asymmetric displacement under loading, where healthy side displaced more. The implanted pelvis plus screw-rod system showed less displacement at implanted side but more at contralateral side in comparison with those without screw-rod system. Background: The modular hemipelvic prosthesis has been used in patient of Type I-IV pelvic tumor with good outcomes, but how to keep the stability between the prosthesis and the residual sacrum is a problem. An additional screw-rod system seems to solve it, but its biomechanical characters are still not well understood, which need experimental evaluation. Methods: Six pelvic specimens were prepared in three conditions (normal intact pelvis, “normal”; the pelvis of left Type I-IV defect and implanted with prosthesis without / with additional screw-rod system, “rod- ” and “rod +”). Compressing biomechanical experiments (50-500N) were performed in these three conditions, respectively. Results: The loadings during the experiments are in accordance with the linear elastic control mode. Under the increasing loading, the implanted pelvises displaced asymmetrically, unlike normal intact pelvis. The vertical displacement of “rod + ” changed significantly, while “rod- ” did not. For both implan ted pelvis, right side displaced less than left side (P values ​​<0.05) .Conclusions: The implanted pelviseded asymmetric displacement under loading, where healthy side displaced more. The implanted pelvis plus screw-rod system showed less displacement at implanted side but more at contralateral side in comparison with those without screw-rod system.