When calculating the production cost of a product, two basic items are taken intoconsideration. The first is the material cost and the second is the labor cost. Materialcosts can be calculated very clearly due to the lack of personal judgment. Because theamount of material used for the production of that product is certain. However, thecalculation of labor costs may not always be as simple and precise. The amount ofproduct produced per unit time directly affects product costs. Therefore, to calculatethe costs correctly, the number of parts produced per unit time must be calculatedcorrectly. At this point, production times per product gain importance. It is notconsidered sufficient to use the values on the chronometer when calculating theproduction times of parts or products. Standard production times are calculated byadding performance and tolerance values to the chronometer values. Workerperformance may vary from person to person. Performance varies depending onworker ability, effort, consistency and working environment. This difference inperformance becomes very important in manual production processes. In automaticproduction processes, this difference is especially important when disassembling andassembly processes of the parts to the automatic machines. Tempo rating systemshave been proposed to standardize worker performances and to minimize thenegative effects of performance differences between individuals. The ClassicalWestinghouse Method (CWM) is one of the common use tempo rating system. In theCWM, the performance of the worker is calculated by using skill, effort,environmental conditions, and consistency values. Evaluations of these four criteriaare obtained by observing by decision-makers. In this study, the fuzzy rule-basedWestinghouse method is developed because decision-makers use linguisticexpressions to evaluate worker performance. The proposed Fuzzy WestinghouseMethod (FWM) is applied to a part of the company that produces automatic cuttingmachines in order to prove its validity. As a result of the study, it is determined thatthe proposed model produces more sensitive values.