The authors consider discrete event models (DEMs) that describe logical system behavior, i.e. mathematical models that represent the set of traces of sequences of events that the system can execute. Examples of DEM formalisms include state machines, Petri nets, and Hoare's communicating sequential processes. All these formalisms are algebras, i.e. a set of models together with operators that combine models to form other models in ways in which real systems are interconnected. A general approach toward constructing model algebras is proposed. The approach covers both deterministic and nondeterministic models, and it can be tailored to recover existing formalisms and to obtain model families. >