Abstract: As a crucial device of steel mills, the draught fan plays a key role in converter dedusting and gas recycling, and thus it is significantly essential to predict the remaining useful life (RUL) of the draught fan. In this paper, a Wiener process-based degradation model is constructed based on vibration data analysis for a draught fan in the Handan steel mill. An analytical expression of the probability density function (PDF) of RUL is derived on the concept of the first hitting time (FHT). A parameter updating scheme is deduced on the basis of the maximum likelihood estimation (MLE) algorithm for the RUL online prediction of the draught fan. Comparative studies with existing models show that the proposed method can predict the RUL of the draught fan in real time with a higher accuracy and less uncertainties.