The instructional value of web-based education systems has been an important area of research in information systems education. This study investigates the effect of various teaching methods on program design learning for students with specific learning styles in web-based education systems. The study takes first-year Computer Science and Information Engineering majors as the sample population. In an earlier study (Hung & Huang, 2008), Bayesian Network analysis of experimental data showed that students with "Input-oriented" and "Perception-oriented" learning styles performed less well on program design tests. Following on these results, this study further analyzes student learning performance in the program design course with a focus on the relationship between teaching methods and learning styles. Students are divided into experimental groups and a control group. The subjects' learning performances are analyzed with the bi-factor variable analysis in SPSS. The results indicate that (a) in learning program design, input-oriented students taught with the diagram method exhibit significantly better learning performance than those using other teaching methods; and (b) in learning program design, perception-oriented students taught with the analogy method exhibit significantly better learning performance than those using other teaching methods. (Contains 5 tables, 3 figures, and 1 footnote.)