Purpose The main purpose of this study was to identify how the accuracy of the tibial rotation reference axes varied in populations with different tibial tubercle locations. We hypothesized that the accuracy of the axes of tibial rotation would be affected by the changes of tibial tubercle locations. Methods Surgical epicondylar axis (SEA), medial third of the patellar tendon (1/3MPT), medial third of the tibial tuberosity (1/3MTT), medial border of the tibial tuberosity (MTT) and Akagi line were drawn. The angle between SEA and horizontal line with the angle between the four tibial rotation axes and the horizontal line was compared by T test. Then, the correlation between TTTG with the angles between the four axes and SEA vertical lines was analyzed. The TTTG was divided into three subgroups (TTTG < 10 mm, 10 mm <= TTTG < 15 mm, TTTG >= 15 mm), then t test was performed for the angles between the vertical lines of the SEA and the four rotation axes of the tibia in each group. Results Among the four tibial rotation axes, only the difference between MTT and the line perpendicular to SEA had no statistical significance (NS.). The four tibial rotational axes were all positively correlated with TTTG (p < 0.001). When TTTG >= 15 mm, Akagi line was 2.5 degrees +/- 6.9 degrees internally rotated to the line perpendicular to SEA, while the 1/3MPT and MTT was 0.9 degrees +/- 5.3 degrees and 1.3 degrees +/- 5.9 degrees externally rotated to the line perpendicular to the SEA when TTTG < 10 mm and 10 mm <= TTTG < 15 mm, respectively. Conclusions MTT showed the best consistency with SEA. TT-TG had a significant positive correlation with all four tibial rotational axes. In patients with TTTG < 10 mm, 10 mm <= TTTG < 15 mm and TTTG >= 15 mm, the 1/3MPT, MTT and Akagi line demonstrated good alignment consistency with SEA, respectively.