There is increasing adoption of robotic surgical technology in Total Knee Arthroplasty. The ROSA® knee system can be used in either image-based mode (using pre-operative calibrated radiographs) or imageless modes (using intra-operative bony registration). The Mako knee system is an image-based system (using a pre-operative CT scan). This study aimed to compare surgical accuracy between the ROSA and Mako systems with specific reference to joint line height, patella height, posterior condylar offset and tibial slope. This was a retrospective review of a prospectively collected data of the initial 50 consecutive ROSA TKAs and the initial 50 consecutive Mako TKAs performed by two high-volume surgeons. To determine the accuracy of component positioning, the immediate post-operative radiograph was reviewed and compared with the immediate pre-operative radiograph with regards to joint line height (JLH), patella height (PH), tibial slope (TS) and posterior condylar offset (PCO). Mean difference between pre- and post-operative radiographs using the ROSA knee system of joint line height was 0.47mm (SD 0.95) posterior condylar offset 0.16mm (SD 0.76), tibial slope 0.9 degrees (SD 1.6) and patella height 0.01 (SD 0.05). Mean difference using the MAKO knee system of joint line height was 0.26 (SD 1.08), posterior condylar offset -0.26mm (SD 0.78), tibial slope 1.8 degrees and patella height 0.03. No significant difference was demonstrated between the accuracy of component positioning of the ROSA or MAKO knee systems. Our study is the first study to compare the accuracy of the ROSA and MAKO knee systems in total knee arthroplasty. Both systems are highly accurate in restoring native posterior condylar offset, joint line height, tibial slope and patella height in TKA with no significant difference demonstrated between the two groups.