A model is a representation of an object, concept or system. A simulation shows the expected working of a system based on a model of the system. A model can be physical or abstract, and both types can be static or dynamic—that is, they stay the same or change with time.

An example of a static physical model is a stick model of a water molecule, with two small hydrogen "balls" stuck with short sticks on either side of the oxygen "ball." This model does not change with time. Another physical model is that of a tank of water with sand, which shows the effect of the wind and the movement of water. In this dynamic model, the sand and water show patterns that depend on the intensity and direction of the wind with time. In most simulation models there is an element of dynamism.

When creating a simulation model, three elements are identified: the parts of the system, the interaction between the parts, and the number and nature of inputs. A model is essentially created for each of these, with crucial aspects considered and minor aspects ignored. Then the simulation model for the whole system is developed.

For example, for a simulation of a factory workflow, one machine can be modeled as an element that takes a certain amount of time to create a particular part, while another machine takes a different amount of time. The time to move parts between machines may be ignored for machines which are close together. The number, the rate, and the time at which the raw material and the work orders come into the factory are modeled. Based on all these, the simulation determines whether the output of the factory meets the demand.

Increasingly, computers are used for simulation. Some simulations can be run with standard simulation programs, and others require special software to be written. The models for the parts, the interaction of the parts, and the inputs are fed to a program. The program then runs the simulation model and delivers the outputs over time, often showing those outputs graphically. With computers, simulations involving thousands or millions of elements and spanning large time intervals can be attempted. Simulations of planetary evolution or advanced military maneuvers are some examples.

Traditionally, the simulation model has been mathematical in nature. Raw material coming into a factory, for example, would be approximated as coming in at fixed intervals. Computers can now do more realistic simulations by using a script that has events taking place that are similar to a real situation or even an exact recording of a real situation.