Supplier selection, which is substantial to make a successful supply chain relationship, is the complex decision making problem with both conflicting multiple criteria and usually several decision makers. This paper extends the existing MCDM (multi-criteria decision making) method to the situation considering lack of information caused by ignorance of knowledge, especially when lack of information happens at all levels including criteria in the hierarchical decision structure. To hurdle the ambiguity and incompleteness of evaluation process, fuzzy linguistic variables are used when the matrices for criteria and alternatives are built. In addition, to deal with lack of information, Demspter-Shafer (DS) evidential theory is applied to the MCDM evaluation scheme. Belief intervals composed of belief function and plausible function are converted to crisp values through the PPT (pignistic probability transformation) method and the quadratic programming with weighted loss objective function. All of the procedures are implemented in detail for a supplier selection example, which shows the newly proposed MCDM evaluation scheme to be applicable and effective.