This research sheds light on one of the most difficult problems especially in designing a translation programme which deals with ambiguity. The programme can be built in a form of categories of words that cannot be adjacent in a sentence. From a semantic point of view, a word certainly has different meanings depending on situations and contexts. It is known that a computer as a machine cannot think as a human being. Thus, our aim is to find a best way of analysing a sentence in a mathematical way, i.e. to find out a programme that uses data bases, calculations, modules that can be linked to each other for specifying the lexical ambiguity.