This paper describes the effects of learning support on simulation-based learning in three learning models: experiment prompting, a hypothesis menu, and step guidance. A simulation learning system was implemented based on these three models, and the differences between simulation-based learning and traditional laboratory learning were explored in the context of physics studies. The effects of the support type on learning performance were also quantified. In second-year junior high school students it was found that the outcome for learning about the basic characteristics of an optical lens was significantly better for simulation-based learning than for laboratory learning. We also investigated the influences of different learning models on the students' abstract reasoning abilities, which showed that the different learning models do not have different effects on individuals with different abstract reasoning abilities. However, we found that students who are better at higher abstract reasoning benefit more from simulation-based learning, and also that the learning results are better for experiment prompting and a hypothesis menu than for step guidance. (Contains 4 figures and 7 tables.)