The introduction of pervasive computing environments enforce new ways of human-machine-interaction. The welldefined interaction interfaces will make place for other, more intuitive ways of interaction. In a pervasive service environment, the system middleware should take care of capturing the users expression of an action intention, solving ambiguousness in this expression, and executing the final pervasive action This article introduces the Pervasive Service Action Query Language (PsaQL), a language to formalize the description of a user intention using composed pervasive services. It presents the next steps of intention treatment in a pervasive service environment: A mathematical model is given, which helps to express the algorithms performing translation of the user intention into an executable action. To implement such algorithms, a suitable object- oriented model representing actions is introduced. In the scope of PERSE, a pervasive service environment developed by our research group, general evaluation metrics for such algorithms are identified, a prototype has been developed and first benchmark results are presented in this article.