With SUVs and minivans accounting for a larger share of the US market in the past decade, rollover accidents have drawn greater attention, leading to more active research from different perspectives. This ranges from investigations for elucidating the basic causes and mechanisms of rollover accidents to studies of more advanced occupant protection measures. As the phenomenon of a rollover accident is longer in duration than frontal, side or rear impacts, it is relatively difficult [1] to simulate such accidents for experimental verification and also for proper evaluation of occupant restraint system performance. In this work, we focused on the trip-over type, which occurs most frequently, and performed simulations to reproduce real-world rollover accidents by combining PC-Crash and FEA. At first, using a simplified full car model, sufficient conditions, such as additional velocity, required for a curb trip-over accident to occur, were derived from energy balance concept based on the same principle as critical sliding velocity (CSV) criterion. Next, as an extension to curb trip-over, a soil trip-over simulation was carried out. Based on rigid body dynamics, PC-Crash software was chosen to make an accident reconstruction analysis of some selected cases chosen from an accident database (NASS-CDS). Through a process of iteraton the overall kinematics of the vehicle movement before and after the crash was captured. The output of this PC-crash simulation was then used as the initial input conditions (i.e., speed, deceleration, etc.) of a detailed finite element analysis. This analysis can simulate more accurately the kinematics of the vehicle and also simultaneously visualize the occupant dynamics starting from the sliding or furrowing (lateral deceleration) phase, through the intermediate tripping and rolling phase to the final contact phase with the ground in a soil trip-over accident. The combination of these two efficient tools helps to reveal the most important parameters in experimental verification for simulating real-world accidents more realistically. Further, virtual testing helps to minimize the number of experiments needed to evaluate occupant restraint systems in new cars.