In the situations in which subjectivity in the interpretation of the data is over randomness, the fuzzy sets provide us with a theoretical framework to solve a wide range of problems in different research areas [12], [13] with a high degree of efficacy and efficiency. In this paper, we are particularly interested in the design of WDM networks with fuzzy demands. With the purpose of solving this problem, we propose two alternative models: a node-segment model in which providers allow violations in the constraints satisfaction, and a segmentpath model in which some coefficients in the constraints are not known accurately. In both cases, in order to solve the optimization problems derived from these situations, we will make use of models which are well known in the area of Fuzzy Mathematical Programming [1], [6], [11]. Because of the high dimension of the auxiliary problems obtained, on these models we will apply solution approaches that provide an original effective way to solve network design problems under uncertainty in the demands.