Opportunistic routing protocols have been proposed as efficient methods to exploit the high node densities in sensor networks to mitigate the effect of varying channel conditions and non-availability of nodes that power down periodically. They work by integrating the network and data link layers so that they can take a joint decision as to the next hop forwarding node based on its availability and suitability as a forwarder. This cross-layer integration makes it harder to optimize the protocol due to the dependencies among the different components of the protocol stack. In this paper, we provide a framework to model opportunistic routing that breaks up the functionality into three separate components and simplifies analysis. The framework is used to model two variants of opportunistic routing and is shown to match well with simulation results. In addition, using the model for performance analysis yields important guidelines for the future design of such protocols.