In the paper we present an ACP-like process algebra which can be used to model both probabilistic and time behaviour of parallel systems. This process algebra is obtained by extension of untimed probabilistic process algebra with constructors that allow the explicit specification of timing aspects. In this paper we concentrate on giving axioms and deduction rules for these constructors. We give two probabilistic process algebras with discrete time. The first one only manipulates with processes that may be initialized within the current time slice or may delay a finite and fixed number of time slices. Later, we add processes whose execution can be postponed for an arbitrary number of time slices.