With the growing number of marketplaces and trading partners in the e-commerce environment, software tools designed to act on behalf of human traders are increasingly being used to automate trading activities. This paper describes a model for constructing trading engines which are capable of concurrently participating in multiple interrelated negotiations with heterogeneous protocols. The behaviour of these trading engines is specified by means of a generic synchronisation construct which enables the incremental composition of complex trading schemes, including a number of well known strategies from the financial trading domain. The construct is augmented with a priority-based scheduling algorithm which selects a set of nodes for negotiation based on their estimated profit, the time remaining and the desired degree of concurrency. The model also supports iterative negotiation, which is essential in any complex trading environment. Copyright © 2005 Inderscience Enterprises Ltd.