This study aims to test the effectiveness of GeoGebra Augmented Reality in improving students' mathematical problem-solving skills. The method used is a quasi-experiment with a posttest-only control group design. The research population is students of the Mathematics Education Study Program at a private university in Indonesia, with a sample of 32 students selected using random sampling techniques (17 students in the experimental group and 15 students in the control group). The study was conducted for 5 weeks with a duration of 2.5 hours per week. Data were collected through mathematical problem-solving tests and analyzed using normality tests (Kolmogorov-Smirnov and Shapiro-Wilk), homogeneity tests (Levene's Test), and independent t-tests after meeting the prerequisites of parametric analysis. The results showed that there was a significant difference between the experimental group using GeoGebra Augmented Reality and the control group using conventional methods (t = 3.790, p < 0.05, Cohen's d = 1.02), with the experimental group experiencing a significant improvement in mathematical problem-solving ability. These results indicate that the application of Augmented Reality technology in mathematics learning can be an effective innovation in improving students' problem-solving skills.