Objective To identify the optimal femoral intramedullary rod insertion depth and direction on the sagittal plane in total knee arthroplasty (TKA) of Chinese osteoarthritis (OA) patients. Methods From January to December 2019, CT data were collected for 85 consecutive entire lower extremity Chinese OA patients. A three‐dimensional method was used to simulate intramedullary rod penetration. The intramedullary rods were inserted toward the anterior (TA), center (TC), and posterior (TP) of the femoral canal, respectively. Four penetration depths of 150, 200, 250, and 300 mm from the joint line were set. The intersection angle was measured between the simulated intramedullary rod and the mechanical axis of the femur (FMA) on the sagittal plane. Results Our study included 85 Chinese OA patients: 46 women, with a mean age of 65.7 ± 8.4 years (range, 51–85 years) and 39 men, with a mean age of 65.6 ± 8.1 years (range, 46–86 years). The intersection angle between the FMA and the femoral anatomical axis was smaller in men, 2.4° ± 1.6° (range, 0°–4.8°), than in women, 3.5° ± 2.3° (range, 0.7º–8.2°), with a significant statistical difference (P < 0.01). In the comparison of the intersection angle between the simulated intramedullary rod and the FMA, there was no statistical difference between TA200 and TC200 in women (P > 0.05). The proportions were up to 91% and 96% of TA200 at 0°–3° and 0°–5° intervals, respectively, but just 63% and 78% in TC200. In TA150, 76% of intersection angles were greater than 5°. Only approximately 60% in TA250 and TA300 were within the 0°–5° interval and 40% were less than 0°. Only 57% of intersection angles in TC150 were in the 0°–3° interval. TC250, TC300, and TP150 were mostly below 0°. In men, there were statistical differences between all groups. All intersection angles were greater than 5° in TA150. TA200 and TA250 were mostly greater than 5° (87% and 59%, respectively) and 72% of intersection angles were within 0°–5° interval in TA300. TC150 had 92% of intersection angles within the 0°–5° interval but only 62% between the 0° and 3° interval. In the TC200, up to 90% and 97% were within 0°–3° and 0°–5° intervals, respectively. TC300, TP150, and TP200 were mostly below 0°. Conclusion We described an innovative method for rapidly, simply, and accurately identifying the sagittal insertion depth and direction of the femoral intramedullary rod in TKA, which can optimize the position of the femoral prosthetic component on the sagittal plane in TKA.