This paper proposes an approach to multi-criteria decision-making which is based on a modified fuzzy TOPSIS method. We assume that the process of determining the ranking of alternatives should depend on subjective preferences for the linguistic values of fuzzy criteria. Preference degrees, which have to be given by a decision-maker, will be used for constructing the vectors \(A^+\), and \(A^-\), representing the positive, and the negative ideal solutions, respectively. The vectors \(A^+\) and \(A^-\) will be described by applying the notion of fuzzy linguistic label. Final ranking will be obtained by evaluating the distance between particular alternatives and the vectors \(A^+\), and \(A^-\). The presented approach will be illustrated by a computational example.