Fuzzy concept lattice generated from fuzzy context (L-context) has been used in a number of fields, such as data mining, decision-making, information retrieval and so on. Recent works have achieved great success. In this paper, given a binary relation, we generated the fuzzy concept lattice. In order to make decision-making much easier, reducing the number of attributes without a large amount of information loss was preferred, a method of computing the similarity matrix was represented as well. We proposed an algorithm for attribute reduction and discussed the time complexity with comparison to the other algorithms afterwards. In the end, the method of computing choice value during decision-making was provided.