Mathematics aims to make students have good problem-solving skills based on the junior high school curriculum. However, the reality in the field is that more than 50% of students cannot solve problem-solving problems. This study aims to (1) Analyze the differences in the improvement of the problem-solving abilities of students who received learning with the GeoGebra-assisted CPS model and the problem-solving skills of students who received conventional learning in terms of overall and early mathematics ability (superior and low); (2) Analyzing the differences in learning interest of students who received GeoGebra-assisted CPS learning and students who received conventional understanding; (3) Analyzing the correlation between problem-solving abilities and students' interest in education. The research design employed was sequential explanatory with a quasi-experimental method. The Instruments were written tests, questionnaires, observation sheets, and interview guidelines. The data obtained were analyzed using the Mann-Whitney U test, t-test, and correlation test. The results showed that: (1) The improvement of the problem-solving abilities of students who received GeoGebra-assisted CPS learning was better than students who received conventional learning, and there was no difference in the problem-solving abilities of the superior early mathematical ability and low early mathematical ability groups; (2) The learning interest of students who receive GeoGebra-assisted CPS learning is better than students who receive conventional learning; (3) There is no correlation between problem-solving ability and student interest in learning. Creative problem-solving learning models assisted by GeoGebra applications can improve students' mathematical problem-solving skills and interest in learning.