The aim of this paper is to present and discuss the application of the epsilon technique to the identification of distributed-parameter systems. After the problem formulation, attention is focused on computational aspects. The numerical results obtained for two typical systems are presented.