This paper utilizes a fuzzy set arithmetic, called fuzzy weighted average (FWA), to evaluate a group of bid tenders (or proposals). Three major criteria and several subcriteria were adopted in the bid proposal evaluation. A rating in terms of natural language was assigned to each criterion, which was then represented by a fuzzy set. The FWA operation was used to combine all ratings in different stages. For each bid a final fuzzy set representing an overall evaluation was obtained. A ranking index model was used to select the best bid. The process was implemented with the FORTRAN-77 language on a microcomputer.