Based on the research of parallel simulation technology for equipment remaining life prediction, the state equations under different degradation states was unified, the construction method of prediction model library and the selection algorithm of predictive model morphological were proposed, and the online model evolution workflow of “model morphology library construction—model selection—data assimilation—model evolution” was designed. Using extended Kalman filter and maximum likelihood estimation algorithm, the online dynamic evolution of parallel simulation model for equipment remaining life prediction was realized. The effectiveness of the model morphology selection algorithm and the model online evolution workflow is verified by simulation experiments, and good simulation results are obtained.