HIV/AIDS is a pandemic that has claimed and continues to threaten the lives of millionsrnaround the globe. In the effort to answer the question "what works?", Voluntary Counselingrnand Testing (VCT) has appeared as a potentially effective intervention to preventrntransmission of HIV by changing sexual behavior, and also in enabling individuals to makerninformed decision on whether or not to be tested. At the present time, VCT is the focalrnpreventive tool in Ethiopia that is also being used as a means of reducing the stigmarnassociated with the disease. The continued efforts to encourage people to come forward forrnvoluntary testing are, however, often hindered due to the stigma and taboo against openrndiscussion about sex and sexually transmittable diseases.In an attempt to address such problems, the objective of this thesis work is to look into thernpossibility of applying the expert system technology to VCT by developing a prototypernexpert system that can imitate the services of a VCT cowlselor. One of the methods used forrnachieving this goal is acquiring knowledge through unstructured intervi,e ws from five expertsrnthat are purposely selected from the domain (VCT counselors). The VCT centers werernselected based on prior acquaintance with the VCT counselors, availability of transportationrnto and from the centers, availability of the counselors to give feed-back and close follow-uprnto the research work, attitude towards the new idea of developing an automated counselor,rnand so on. Secondary sources of knowledge, such as standard VCT guidelines from UNWHOrnand Centre for Disease Control were also used to gather knowledge about the domain.rnThis knowledge is then represented into a form that can be understood by a computer usingrnproduction rules and mapping these rules into an expert system shell called KnowledgernProGold Version 3.0. The resulting prototype, Voluntary Counseling Expert System onrnHIV/AIDS (VCE-AIDS), includes a user-interface that serves to display questions, acceptrninput from the user, and present the resulting decision of the system.VCE-AIDS has been tested for performance on IS users and achieved quite an encouragingrnresult of about 90% accuracy (86.6%). VCE-AIDS is, however, limited to the pre-HIV testingrnsession of the counseling process and is recommended to be developed into a full fledgedrnsystem that includes post-HIV -test counseling.