Artificial Neural Network (ANN) models have been widely applied to many hydrologic systems such as hydrologic modeling, hydro-power and reservoir operations, water demand, and some environmental variables (e.g. sediment, salinity, pH, and temperature). The majority of these applications are based on training the ANN model by minimizing the fitting error. Although ANN models provide excellent function fitting capability, these “best-fitting” solutions provide little information about the estimation uncertainty. Recent development of sequential Monte Carlo (SMC) simulation techniques provides a class of algorithms to address the uncertainty of estimates. The advantage of using SMC techniques is the flexibility of processing new observations in the ANN model in near real-time and is capable of dealing with nonlinear models and non-Gaussian error characteristics. In the presentation, the development of ANN models in the precipitation monitoring from the remote sensing information and in streamflow forecasting will be discussed. The uncertainty model estimates based on SMC techniques will also be addressed.