With the advance of science and medicine, life expectancy of human beings becomes increasingly higher. Never the less, human bones and joints very often could be impaired and cause pain due to degeneration, osteoporosis, and long-term use. With the advance in science and medical engineering, artificial joint replacement can greatly relieve pain and improve the functionality of the joint in cases of severe joint diseases. X-ray Computed Tomography (abbreviated CT) is a medical imaging test that employs x-ray to penetrate human body and x-ray sensors positioned at the opposite side of the body to receive the x-ray. The x-ray images are then processed by computer to create tomography, but errors in measuring accuracy none the less could cause distortion to the computed tomography. Previous concept of designing artificial joint component has been existent in simulated situation, and there are many insurmountable technical difficulties in the practice of reconstruction which results in design flaw in the geometric shape of artificial joint body and increases the risk of failure. This study uses empirical method to determine the shape and size that is closest to human bone, and aims to achieve accuracy error compensation by means of correction mode and consequently improve the accuracy of reconstructed artificial joint and lower the risk of failure in artificial joint arthroplasty.