This paper aims at investigating the cause-and-effect relationship among classroom community, self-efficacy and academic ability in a group of four-year college students. The 167 participants took a fifteen-week course of English classes and completed a questionnaire on self-efficacy and classroom community. T-tests, ANOVA and multiple regression analyses were conducted to examine the relationships among all variables. The findings showed that, first, confidence, connectedness and learning were statistically significant predictors when comparing subcategories of self-efficacy and classroom community between gender. Second, among subcategories of self-efficacy and classroom community, confidence and learning were statistically meaningful indicators of academic ability. Third, all the subcategories in self-efficacy and connectedness in classroom community were statistically meaningful indicators of academic ability. Finally, multiple regression analyses predicted that confidence and interdependence have a cause-and-effect relationship in college students’ academic ability. In conclusion, the casual relationship among confidence and interdependence and academic ability was found to be definite. Implications for the practical class and further effective studies are suggested.