Sign language is a visual-gestural language mainly used by hearing-impaired people to communicate with each other. As in oral language, sign language is not universal; it varies according to the country, or even according to regions. Ethiopian Sign Language (ESL) is the sign language of the deaf in Ethiopia. Hearing-impaired people usually face communication problems when they want to communicate with hearing people without signing skill. At present, sign language interpreters are used to remove the language barriers between people who are deaf and use sign language and people who can hear and speak. However, cost, availability and privacy issues make this an awkward solution at best. A computer application which translates written text into sign language, would better bridge the communication gap between deaf and hearing worlds. This thesis was therefore to contribute in this area by developing translator software, which automatically translates Amharic text into Ethiopian Sign Language (ESL). The developed translator is referred in this thesis as AmESL-T. The system is composed of a text-analysis, mapping and sign animation (using 3D avatar) module. The input for the system is Amharic word, letter and number. The output of the system is avatar animation of Ethiopian sign language. The signs are represented by means of VGuido (the eSIGN 3D avatar) animations. For creating gesture animations, eSIGN editor is used. Amharic has complex morphology which makes complete listing of all words in the database impossible. Hence, a morphological analyzer that reduces morphological variants of a word into a common form is developed and integrated to AmESL-T. AmESL-T is constructed as a web application to allow usage to large population. Using the translator, it is possible to translate Amharic texts into gestures that a hearing impaired can understand. The tool is useful in enabling people who don’t know sign language to communicate with deaf individuals. The hearing-impaired can also utilize the software to develop written language skills. Moreover, it may be used as a teaching tool for Ethiopian sign language. AmESL-T was evaluated in terms of quality of signing and subjective opinions. Quality of signing was measured based on intelligibility and acceptability of signs. The evaluation results showed that the quality of signing and the translation accuracy are acceptable, and it satisfies users' need. For better translation results, further research areas are also pointed out.