We propose a model delineating the role of control, value, and cognitive disequilibrium in elementary students' experience of emotions during mathematics problem solving. We tested this model across 2 studies. In Study 1, using an explanatory mixed-methods design, 136 students from Grades 3 to 6 worked on a complex mathematics problem appropriate for their grade level. A think-aloud protocol was used to capture cognitive processes, and trend analyses were applied to students' transcriptions to assess convergence or divergence of the quantitative results and to provide a richer account of students' experiences. Results revealed that cognitive disequilibrium mediated relations between control and emotions but not value. In addition, curiosity and frustration predicted enactment and metacognitive strategies, which directly predicted mathematics problem-solving achievement. Trend analyses revealed that confusion following a failed attempt at impasse resolution led to frustration for 33% of the instances of a failed attempt. Interestingly, 35% of those instances resulted in students seeking help, and 32% resulted in students trying again or moving on. Additionally, trend analyses provided evidence that curiosity following surprise was a function of high resolution expectancy, whereas confusion following surprise was because of low resolution expectancy or high complexity. In Study 2, a new sample of 80 Grade 5 students completed a multiday complex mathematics problem and self-reported their emotions and cognitive appraisals of control and value (pretest and posttest). Results revealed reciprocal relations between control and curiosity and control and frustration. No reciprocal effects were found for value. Theoretical and practical implications are discussed.