An online soot load estimation method based on mass balance was developed through establishing the predicting models of NOx and soot emissions from engine and the model of catalytic soot oxidation reaction of passive regeneration in diesel particulate filter(DPF). The verification of the emissions models were carried out by the new European driving cycle(NEDC) test. The results show that the prediction errors of the NOx and soot emissions models are 5.1% and 3.9% respectively. A DPF soot loading test is also conducted by on-vehicle real world driving. It is shown that the maximum error of the developed estimation method is 0.48 g/L and the average error is 0.17 g/L

achieving an average online estimation error of 2.1% during the whole testing phase. This indicates that the method has a considerable potential in performing the DPF thermal regeneration in a reasonable and timely manner.