Abstract: In order to rapidly find out causes of failure of underground water pump, a fault diagnosis model of underground water pump based on fuzzy Petri net and condition monitoring was established. Firstly, vibration signal of the water pump was measured by the condition monitoring system of underground drainage equipment, training was carried out on the water pump fault samples after vibration analysis. Then, on the structure of fuzzy Petri net model of water pump fault diagnosis, BP algorithm of neural network was introduced to train parameters such as weight values, threshold values and credibility. The results of instances analysis show that the model can be used to find out the causes of pump failure accurately, and has good accuracy, rapidity and adaptability.