本文探讨大坝监测资料的动态分析问题,包括贝叶斯线性动态模型的建立,递推解法,以及初始先验估计和状态噪声的取值等.从二个实例的数据看,成果是合理的,比现用的最小二乘回归模型要好,预测精度高,能实现在线监控,表明大坝监测分析中应用贝叶斯方法是可行的,值得推广应用. This paper discusses the dynamic analysis of dam monitoring data, including the establishment of Bayesian linear dynamic model, recursive solution, as well as the initial a priori estimates and the value of state noise, etc. From the data of two examples, the results are reasonable , Which is better than the current least-squares regression model. The prediction accuracy is high and on-line monitoring can be realized. It is feasible to apply Bayesian method in dam monitoring and analysis, which is worth to be popularized and applied.