Supplier selection process is challenged by compiling information from decision-makers who often have different roles within organisations, with competing priorities. This difference in decision-makers' perspectives can introduce bias in the supplier selection process. Evaluation of suppliers is also challenged by decision-makers' inability to provide a precise numerical value reflecting their choices. To address these issues, this paper develops a new supplier selection model that enables a systematic process for aggregating preferences of non-homogeneous decision-makers who use linguistic labels to express their preferences. The model also considers two critical pieces of information: Reliability of criteria and importance of assessment values (with respect to their ordering). A case study is used to demonstrate application of the proposed model. The model and its procedure can help managers select the most appropriate supplier that satisfies the organisation's requirements.