A model of multicriteria decision making is developed taking into account the reliability of the data obtained. To formalize the information containing the data and assess their reliability, Z-numbers are used, the definition of which was given by Lotfie Zadeh in 2011. Most of the well-known decision models based on Z-numbers are limited by the assumption of a probabilistic assessment of the reliability of the data, which significantly narrows the scope of these models. This article partially removes the restrictive requirements when working with Z-numbers. For components of Z-numbers, aggregate indicators are calculated using a-cuts, based on which the similarity indicator between Z-numbers is determined. Choosing the best alternative is based on the minimum indicator of similarity with the ideal alternative. A numerical example is presented that shows the operation of the model and its effectiveness under conditions of multi-criteria selection.