Job performance measurement is of critical importance to any organization.s health. It is important not only to recognize and reward good performance, but also to groom future leaders. Developing effective assessment techniques that are valid, effective and fair is an ongoing challenge. Assessing factual knowledge using multiple-choice test batteries relatively inexpensive and tends to be commonly used. Hands-on assessment is the most effective in assessing task proficiency but is very resource intensive and expensive. Computer-based simulations provide an alternative where users can be assessed in the context of skill application under controlled conditions. However, simulations are expensive to produce and maintain. Validated guidelines and methodologies are needed to help organizations develop effective assessment simulations. In this paper we present a standard, prescriptive methodology for developing simulations for job performance assessment. We then describe a performance assessment simulation for Light-Wheeled Vehicle Maintenance constructed according to this methodology. This simulation includes automated assessment methods that borrow heavily from existing work in intelligent tutoring systems. Finally, we discuss future research directions based on the results of this initial methodology and assessment.