Simulation actually presupposes theory: you need accurate equations (and experimental values too) before a simulation can be made in the first place. Of course, there can be interaction between theory and simulation (e.g. testing whether a simulation with a set of equation yield observables that agree with our experimental observartions), but they are still different.
this might be counter-intuitive because no experiment is necessarily done. Both theory and simulation can be done (in principle) with pen and paper, or more realistically on computers.