The commercial vehicle is widely used in the overland transport. A prediction is given on the 9th annual China automotive industry forum that the number of the global commercial vehicles will reach eight million by the year of 2016. However, since the distance between its gravity centre and the ground is larger than that of the passenger vehicle, considering its comparatively short wheelbase, the rollover accident, which is fatal to the drivers and always makes enormous loss of merchandises, easily occurs in the case of commercial vehicles. As the number of the commercial vehicle is increasing fast, the accidents will occur more frequently, the losses will be increasingly enormous. To solve the problem, many researches about rollover early warning systems have been done. In most cases, it is assumed that the references of the vehicle are given. Nevertheless, most of these references cannot be easily measured, particularly the gravity centre which is always changing with the quantities of the goods. In order to give a more practical solution, a state estimator based on the AKF is designed. This estimator calculates the position of C.G point firstly using the normal braking condition.