Model-based machinery diagnostic and prognostic techniques depend upon high-quality mathematical models of the plant. Modeling uncertainties and errors decrease system sensitivity to faults and decrease the accuracy of failure prognoses. However, the behavior of many physical systems changes slowly over time as the system ages. These changes may be perfectly normal and not indicative of impending fail-ures; however, if a static a priori model is used, modeling errors may increase over time, which can ad-versely effect health monitoring system performance. Clearly, one method to address this problem is to employ a model that adapts to system changes over time. The risk in using data-driven models that learn online to support model-based diagnostics is that the models may ``adapt'' to a system failure, thus ren-dering it undetectable by the diagnostic algorithms. An inherent trade-off exists between accurately track-ing normal variations in system dynamics and potentially obscuring slow-onset failures by adapting to failure precursors that would be evident using static models. Barron Associates, Inc. and the University of Virginia propose an innovative solution that brings together Barron Associates' proven model-based diagnostic and prognostic algorithms with adaptive system identi-fication algorithms enhanced specifically for health monitoring applications that would benefit from online learning.