The EPA is required by Title III of the 1990 Clean Air Act Amendments to regulate the emissions of 189 hazardous air pollutants (including trace metals) from industrial facilities. These facilities may be required to apply technology-based standards such as maximum achievable control technology (MACT) to their emitting processes. A new legislation will be implemented by specific rules and regulations promulgated by federal (EPA), state, and local agencies. Once rules are promulgated, facilities may have to conduct compliance tests and trial burns to demonstrate compliance with emission limits. The process of planning and performing such tests is formidable due to the planning time requirement and the high cost. This paper presents a computer modeling approach to predict metals behavior in coal-fired combustion systems. This treatment is not intended to be a substitute for actual compliance tests but can serve as a guide to help planners, engineers, and facility owners/operators predict the levels of metals emissions from their facilities, and to gain in insight into the impact of operating conditions on metals behavior. To test the predictive applicability of the modeling approach, predictions of the overall partitioning of metals in a pulverized bituminous coal-fired combustor were compared to measured data from full-scale facilities in the Netherlands. In general, the comparison between predicted and measured values is favorable. This indicates that using the modeling approach described in this paper can be effective in planning tests, optimizing operating conditions, and determining the effectiveness of air pollution control equipment prior to conducting actual compliance tests.