In this study, path analysis modeling is applied to examine the relationships among e-learning systems, self-efficacy, and students' perceived learning outcomes in the context of university online courses. Independent variables included in the study are e-learning system quality, information quality, computer self-efficacy, system-use, self-regulated learning behavior, and user satisfaction as potential determinants of online learning outcomes. A total of 674 valid unduplicated responses from students who have completed at least one online course at a university in the Midwest were used to fit the path analysis model. The results indicated that system quality, information quality, and computer self-efficacy all affected system use, user satisfaction, and self-managed learning behavior. But our data failed to show the two relationships: system quality is not positively related to system use; and computer self-efficacy is not positively related to user satisfaction. Two mediating variables (user satisfaction and self-regulated learning behavior) affected students’ e-learning outcomes. But our model failed to see the effect of systems use on e-learning outcomes.