To perform real-time risk assessment for a multi-component system like the steam turbine in a nuclear power plant, the risk of the system must be predicted by taking into account multiple degradation processes and the relevant condition monitoring data. In many situations, multivariate condition monitoring data are collected asynchronously due to different sampling rates and/or variable condition monitoring schemes used in sensor networks. This raises a challenging problem in probabilistic risk assessment for such complex systems. In this paper, a Functional Principal Component Analysis (FPCA) method is developed to overcome this challenge. This method enables the prediction of remaining useful life of each component in the system and the prediction of real-time system risk based on multivariate asynchronous condition monitoring data. A numerical example is provided to demonstrate the application of the proposed method.