In condition-based maintenance practice, one of the primary concerns of maintenance managers is how long a monitored item can still survive given condition monitoring information up to date. Once such a model of the residual life is constructed, sequential maintenance decision-making model is readily set up to aid decision-making. The paper reports a model to predict the residual life distribution of a monitored item based on the measured condition monitoring history information up to date. The residual life of a monitored item can't be described directly by the measured condition monitoring information, but is assumed to correlate with it stochastically. The stochastic filtering theory is applied to establish the relationship between the unobservable residual life of a monitored item and available condition monitoring history information up to date. The model is relevant to a large class of condition monitoring techniques currently used in industry. Not only does the model make the most of all available condition monitoring history information up to date, but also the modeling process is dynamic, and whenever a new piece of information becomes available, the conditional distribution of the residual life will be updated. Method of estimating the parameters in the model is also discussed. A case example is presented to illustrate the modeling ideas.