The paper presents an approach to QoS selection of webservices. It introduces a theoretical frame and respective applicableselection procedures, in which the solution accounts for the uncertainty ofthe existing metrics data and client preference. Methods of two generalcases are revealed in detail. The first method considers a selection basedon one service quality. The second method assesses the integrated QoS ofinteresting properties. The theoretical analysis is validated throughexperimental investigation of real data of services’ quality metrics.ACM Computing Classification System (1998): H.3.7, K.3.1.* The author acknowledges the financial support of the National Science Fund, BulgarianMinistry of Education and Science, under project grant DN 02/11 and the Science Fund ofthe St. Kliment Ohridski University of Sofia under project grant No 80-10-196/24.04.2017.