In the flexible production line evaluation process, fuzzy multiple attribute decision making is frequently employed to resist the ambiguous and incomplete information. And large amount of time and cost have been paid to gather the information so as to construct the decision matrix. Besides, the computational efforts spent in deriving the ranking orders of the alternatives grow rapidly as the number of attributes increases. However, in a typical fuzzy decision making problem with multi-attributes, not all of the attributes are necessary to reach the final decision. Redundant attributes are inevitable, especially when the problem is large-scaled and complicated. In this paper, we will introduce the simplification theory and techniques in fuzzy multiple attribute decision making problems, which incur great costs especially in the production line evaluation process. To solve this problem, two attribute reduction algorithms are proposed. The purpose of the algorithms is to attain all of the simplified order-preserving attribute subsets and the order-preserving attribute core subset of the corresponding multiple attribute decision making problem. Then one of the simplest order-preserving attribute subsets will be employed as a substitute of the original attribute set. Furthermore, via comparison, dissimilarities between the attribute reduction in decision making and the knowledge reduction in rough set theory are discovered. Finally, a case study concerning production line evaluation is analyzed in depth to demonstrate our approach, and also a sensitivity analysis is conducted to reflect the robustness of each attribute reduct.