Decision-making in enterprises is often connected with selecting the best solution on the basis of many criteria describing the analyzed problem. From this point of view, it can be called a multi–criterial decision–making problem. The article presents the use of a chosen clustering method – the k-means method – in the selection of materials for the SAW method process. Clusters were divided into two, based on the two different ways of choosing their initial centers. The two options were compared, and finally the clusters created on the basis of the chosen division were characterized in detail. The selected approach proved useful as decision-making support for purchasing materials in the welding industry.