Nowadays, with the development of the Internet of Things (IoT), the relationship between sensor manufacturing technology and wireless network communication technology is getting closer. It is a great direction that diagnosing motor fault using the sensors with information perception, data processing, and wireless communication capabilities. To reduce the memory requirements, and improve the accuracy and stability of the fault diagnosis, we propose a motor fault diagnosis method based on the industrial wireless sensor network. Our proposed method includes an early warning method based on Bloom filter and a fault diagnosis based method based on decision tree-Bayesian network. Simulation results show that our proposed early warning method can reduce the memory requirements compare to the tradinational early warning method based on a hash table. In addition, simulation results also show that our proposed fault diagnosis method can achieve higher diagnostic accuracy compared to the fault diagnosis method based on traditional Bayesian network and diagnostic Bayesian networks. Moroever, we evaluate our proposed method by experiments. Experimental results show that our proposed method can effectively solve the problem of information data uncertainty in the field of motor fault diagnosis, which verifies that our proposed motor fault diagnosis method can achieve high stability.