The work presented in this thesis lies in the general field of natural language processing and is concerned with machine translation specifically. Machine translation (MT) is considered to be one of the important application areas where the use of computational linguistic techniques are involved :i.e the analysis and generation of natural language sentences. While there are several known approaches and strategies in MT the strategy chosen for this work is that based on the interlingual technique, which represents a popular method for MT when several languages are involved, and this technique is coupled with the interactive and post-editing approaches. The designed system started with previous existing work involving English and French languages, and then the inclusion of Arabic in the system which is the main aim of this work. The result is a system capable of handling MT between any pair of languages in both directions for simple sentences in a chosen narrow domain.The system is called FEAST: an acronym for French English Arabic System Translator. Although the system in its present form is limited to some simple structures of the sentences in the three languages involved, it is aimed to be flexible in the sense that more types may be added by expanding the acceptable grammar rules, dictionaries, and indeed other languages may be included. FEAST is based on the syntatical and semantical analysis of sentences involved capable of solving some of the situations where ambiguities and irregularities at the level of dictionary entries is concerned. FEAST is implemented on an IBM PC compatible machine using Turbo Prolog 2.0, under MS DOS version 5 with Arabic support.