The synchronous operation of the digital twin model of nuclear power and the actual unit is an important guarantee for the safe operation of nuclear power. Taking the digital twin model of nuclear power steam system as the research object, the proxy model of steam system was constructed by constructing a five-layer fully connected neural network, and then the intelligent optimization algorithm was used to optimize the parameters of the proxy model to realize the automatic synchronization of the digital twin model with the actual unit, and finally an improvement scheme of the intelligent optimization algorithm was proposed and its application was verified. The results show that the proxy model can reach 99.99% of the accuracy of the digital twin model, the call speed is increased by 5 times, the optimization results of the intelligent optimization algorithm are better than the manual debugging results, the improved scheme can effectively improve the optimization accuracy, and the automatic synchronization effect is remarkable.