The water-surface takeoff of a trans-medium vehicle involves a transition from the high-density water phase to the low-density air phase, presenting challenges in terms of high hydrodynamic resistance and pitch instability. Accurate prediction of takeoff performance is critical for vehicle design. In this study, unsteady Reynolds-averaged Navier–Stokes (URANS) simulations are conducted using the Realizable k-ε turbulence model and the volume-of-fluid (VOF) method for free-surface capturing [1]. An overset mesh approach is employed to resolve 6-DOF motions under self-propulsion conditions. A thrust model derived from wind tunnel measurements of the ducted propeller is implemented as a velocity-dependent function. Takeoff is defined as the moment when the hull completely loses contact with the water surface (zero wetted area). Self-propulsion simulations reveal that thrust applied at the propeller location (above the center of gravity) induces a nose-down moment, causing excessive pitch (up to 25°) and preventing acceleration beyond 5 m/s. In contrast, with thrust aligned with the center of gravity, the vehicle successfully accelerates and becomes fully airborne. The typical takeoff sequence is illustrated in Fig. 1. Nine geometric modifications are evaluated under identical thrust input. A 1 cm increase in forward step height achieves takeoff at 17.3 m/s. A V-shaped bottom (2 cm width, 15° deadrise) achieves 16.1 m/s. The surface-piercing hydrofoil configuration achieves the lowest takeoff speed at 9.9 m/s (compared to 20.6 m/s for the baseline), demonstrating a significant improvement in thrust utilization efficiency. Longitudinal moment balance is the key factor governing takeoff success. Thrust misalignment with the center of gravity severely impairs acceleration, while geometric modifications, particularly surface-piercing hydrofoils, substantially improve takeoff efficiency.