This study investigated how motivational strategies predict ninth graders' mathematical problem-solving achievements. The students were grouped based on their school types as students who have high and moderate academic achievements. Quantitative data were collected by motivational strategies for Learning questionnaire and routine/nonroutine mathematical problem-solving tests. Data were analyzed by multiple regression, Pearson's r, and Welch's t-test. Results showed that ninth graders with high and moderate academic achievements need to improve their problem-solving skills. A statistically substantive difference in routine problem-solving success and a statistically moderate difference in non-routine problem-solving success between the two groups were found. Self-efficacy was the strongest significant predictor for both the problem-solving achievements (routine/nonroutine) of students with moderate academic success. Cognitive strategy use was not a significant predictor for routine problem-solving achievements of both groups. Test anxiety predicted both groups' routine problem-solving achievements negatively. Text anxiety also predicted non-routine problem-solving achievements of students with moderate academic success negatively. All variables predicted non-routine problem-solving achievements of students with moderate academic success. The analysis revealed no correlation between routine and non-routine problem-solving achievements of students with high academic success. The analysis of students with moderate academic success revealed that there is a statistically moderate and positive correlation between routine and non-routine problem-solving achievements.