Abstract: Process algebra includes a series of formal languages that are suitable to describe concurrent and communication systems including Web services. Nowadays, although process algebra has been effectively exploited for modeling and verification of Web service composition, only functional aspects of Web service composition have been addressed, and non-functional aspects have been ignored due to the fact that process algebra lacks for capability of modeling them. Since quality of service (QoS) of Web services composition is an important non-functional aspect and may include many quality attributes, such as resource, time, fee, etc, an abstract concept, which is cost, is proposed to model quality attribute. CCS(calculus of communicating systems) that is a classical process algebra with this abstract concept is extended, and a new process algebra called PPA(priced process algebra) is proposed. In PPA, actions and states of a process are associated with a priced function respectively. By this way, each action can be tagged with price and each state can be coupled with cost. Syntax and semantics of PPA are then presented. Cost weak bi-simulation of PPA is defined and relationship of it and weak bi-simulation of CCS is then presented. It is proved that PPA extends CCS with cost modeling capability. And an algorithm is proposed to construct cost state space that is used to select Web services composition with optimal cost. Experimental results show that PPA can not only model functional aspects but also non-functional aspects of Web service composition.