Agricultural decision making at the production and policy levels involves processes which can be enhanced by modern computer technologires. Crop and soil models can be used to estimate the impact of different p:roduction strategies at a regional level, and thus provide supplemental lnformation to help decision makers select management practices in accordance with their p.lanning objectives. '.rhls paper describes the structure and implementation of an agricultural decision support system named AEGIS ·• Agricultural and Environmental Get>graphic Information System - designed to evaluate production strategies and to facilitate the creation of a reg:Lonal agricultural plan. AEGIS links s:Lmulation results from two IBSNAT crot? growth models (BEANGRO and CERES-RICE) and a soil erosion model (USLE) to a geographic lnformation·system (pcARC/INFO) and an expert system. A menu-driven interface ls developed to help users with limited expertise in computers interact with AEGIS. Three areas in Puerto Rico were us·ed to develop a prototype of this syste1n. An example ls presented to demonstrate the potential use of thla system as a planning tool for regional policy makers. The role of the models, the organization of the databases, th,~ design of the user interface and lthe expert system are discussed. Thh model-based system demonstrates the effectiveness of modern computer technologies in the processes of agricultural planning, information management, and agrotechnology transfer in Ca1rlbbean countries.