Computer simulations are an often applied and promising form of CAL. A main characteristic of computer simulations is that the domain knowledge is represented in amodel. This model contains all necessary information to calculate the behaviour of the simulation in terms of variables and parameters and a set of rules or constraints which determine the changes to the values of the variables. In order to increase the learning effects of computer simulations additional support and guidance should be offered to the learner. This means that simulations should be embedded into a supportive environment, which we will call an Intelligent Simulation Learning Environment (ISLE). One of the basic components of the ISLE should be a formalised representation of the domain. In this paper the structure of this domain representation and its authoring will be discussed. It is argued that the simulation model is a necessary but certainly not sufficient source of information for building a domain representation for an ISLE. Besides a behavioural description as given by the simulation model (the term runnable model will be used) also a cognitive description of the domain is needed. This cognitive model forms the basis for a number of functions to be performed in an ISLE, like diagnosis, instruction and support. The current paper presents a framework which can be used to formalise the cognitive model. In particular the component of the cognitive model which contains a conceptual representation of the domain, the conceptual model will be discussed. An important element of the framework presented is a relation typology which describes the interrelationships between relations that are used for the construction of a cognitive model. This typology will be an important knowledge source for an ISLE and can support the author with constructing the conceptual model.