The ventilation system conveys fresh air into laneway, dilutes the noxious gases, and keeps the working environment safe and reliable, how to evaluate the performance or reliability of ventilation system has been one of the important issues in coal mine safety research field. The reliability of underground coal face ventilation was discussed in this paper based on multi-sensor system data fusion approach. A partial-fusion architecture were built and a new basic probability assignment(BPA)based on distance function approach was brought forward. The final case study showed a good effectiveness and feasibility of the data fusion approach.