The purpose of this research was to study the effects of selected characteristics of a web-aided instructional simulation on students' conceptual development, problem solving and transfer. A two-pronged research study was conducted using students enrolled in a beginning meteorology course at Iowa State University. For the experimental design, the performances of three groups (”using a simulation With-log” group, ”using a simulation without-log group” and ”control” group) were compared on posttest scores and a weather forecasting activity. No statistically significant differences (P＜0.05) were found between the groups on these measures. However, follow-up interview data obtained from five diverse students in the treatment group showed that the simulation with authentic situations, multiple representations, and the capability of reviewing previous actions supported science learning. For the student who was able to readily take advantage of the information contained in the log and graphs, the simulation exercises were quickly completed and a sufficient understanding of the concept was developed to transfer the newly-gained knowledge to a new situation. For others the simulation was less effective. Although multiple representations can be used by the students to construct knowledge, those who make connections between representations can get better understandings of the simulated phenomena. According to Goos's and Galbraith's model of problem solving (1996), only one interviewee had the features of all three stages. This student also demonstrated a better understanding of how to transfer newly gained knowledge to weather forecast exercises. The other four interviewees, who did not have the features of all stages, need teachers' facilitation to build their problem-solving skills in order to make optimal learning using the instructional simulation developed by this study.