Introduction: Simulation training is an increasingly used method to train medical students in the use of ultrasound guidance for vascular access positioning. Although very efficient for basic training, commercial simulators for vascular access do not reproduce real-life conditions. We developed a biological training model, using porcine liver, and compared it with an existing commercial model. Methods: Whole porcine livers were used by perfusing the portal vein system after inferior vena cava clamping. Thirty-three practitioners accustomed to ultrasound-guided procedure were enrolled to perform an ultrasound-guided vascular procedure on both biological and commercial models. Procedure duration was recorded and 10-point scales were used to compare the 2 models regarding image quality, procedure feeling, and similarity with the real-life procedure. Results: Participants reported a better image quality with the biological model (8.8 ± 1 vs. 7.7 ± 2, P = 0.007) as well as a significant difference in the procedure feeling (8.0 ± 1 vs. 6.9 ± 1.9, P = 0.002). Real-life likeness was significantly better for the biological model (8.4 ± 1.1 vs. 4.5 ± 6, P < 0.0001). Procedure duration was almost 3 times longer using the biological model than the commercial model (209.6 ± 189.0 vs. 59.8 ± 50.1, P < 0.0001). Conclusions: This study validates our biological model of porcine's liver as an interesting training model, allowing closer real-life perception than its commercial counterpart. This model could complement and enhance simulation learning.