Kalman filter is a linear model based on state equation, which inputs and outputs the measured data, so as to obtain the optimal equipment parameter state prediction method. In practical application, considering the influence of noise and disturbance on the measurement results, the optimized estimation can be regarded as a filtering process. By analyzing the measured noise and using Kalman filter, a group of data containing measured noise can be predicted. This algorithm is easy to program and can update and process the field data in real time. So the Kalman filter is a kind of equipment fault prediction method that is easy to implement and combine with other algorithms. Through the discussion of the key methods of Kalman filter and the analysis of examples, this paper shows that Kalman filter has a good application prospect in equipment fault prediction.