Artificial knee arthroplasty, as the most effective method for the treatment of end-stage joint diseases such as knee osteoarthritis and rheumatoid arthritis, is widely used in the field of joint surgery. At present, Finite element analysis (FEA) has been widely used in artificial knee replacement biomechanical research. This review presents the current hotspots for the application of FEA in the field of artificial knee replacement by reviewing the existing research literature and, by comparison, summarizes guidance and recommendations for artificial knee replacement surgery. We believe that lower contact stress can produce less wear and complications when components move against each other, in the process of total knee arthroplasty (TKA), mobile-bearing prostheses reduce the contact surface stress of the tibial-femoral joint compared with fixed-bearing prostheses, thus reducing the wear of the polyethylene insert. Compared with mechanical alignment, kinematic alignment reduces the maximum stress and maximum strain of the femoral component and polyethylene insert in TKA, and the lower stress reduces the wear of the joint contact surface and prolongs the life of the prosthesis. In the unicompartmental knee arthroplasty (UKA), the femoral and tibial components of mobile-bearing prostheses have better conformity, which can reduce the wear of the components, while local stress concentration caused by excessive overconformity of fixed-bearing prostheses should be avoided in UKA to prevent accelerated wear of the components, the mobile-bearing prosthesis maintained in the coronal position from 4° varus to 4° valgus and the fixed-bearing prosthesis implanted in the neutral position (0°) are recommended. In revision total knee arthroplasty (RTKA), the stem implant design should maintain the best balance between preserving bone and reducing stress around the prosthesis after implantation. Compared with cemented stems, cementless press-fit femoral stems show higher fretting, for tibial plateau bone defects, porous metal blocks are more effective in stress dispersion. Finally, compared with traditional mechanical research methods, FEA methods can yield relatively accurate simulations, which could compensate for the deficiencies of traditional mechanics in knee joint research. Thus, FEA has great potential for applications in the field of medicine.