Public policy and program implementation has come to be regarded as a significant problem area in the governance process. Research has provided insights but falls short of totally satisfactory prescriptions for operational success. The literature on policy and program implementation reflects a dichotomy of means between centralized control and delegation of substantial discretionary authority. The resulting theory leaves a gap with practice. Scenario writing is one of the techniques available to practitioners that might be employed to assist in the execution of their responsibilities. Scenarios can be useful tools, but their preparation is costly and time consuming. It was hypothesized that computer modeling techniques and artificial intelligence could be applied to scenario generation to create an effective, practical instrument to permit wider and more effective use of scenarios for planning and management. A computer supported procedure is presented for generating scenarios which describe alternative sequences of future events for a given situation and proposed policy. The generator design reflects a three-way compromise between processing flexibility, data-base structure, and user workload requirements. This prototype generator was subjected to exploratory trials. The lessons learned highlight some potentially valuable program improvements, the importance of focusing the scenario at a level useful to the reader, and the need for a common set of definitions. / Ph. D.