The impact of emotions on learning is an important topic. Yet, compared to cognition and motivation constructs, academic emotion has not been widely explored. Therefore, the purpose of this study was to: (a) test the fit between empirically observed data and the control-value model of positive/negative emotions proposed by the authors; (b) analyze the effects of environmental goal structures on control-value beliefs and the indirect effects on academic emotions through control-value beliefs; and (c) confirm the proposition that value constructs has greater effect on motivational process than expectancy constructs of the expectancy-value theory. Participants were 635 junior high school students. The instruments used in this study included the Academic Emotion Scale, the Environmental Goal Structure Scale, and the Control-value Belief Scale. Observed data were analyzed by SEM and model comparison approach. Results of this study were as follows: (a) There is adequate fit between the observed data and the theoretical models; (b) The environmental goal structures have direct effects on control-value beliefs, and have indirect effects on achievement emotions through control-value beliefs; (c) Task value has greater effect on positive emotions than self-efficacy; however, both of them have no difference in effects on negative emotions. Implications for theory, instructional intervention and future research are discussed.