This study aims to analyze the effect of learning based on concreteness fading and multi-representation on prospective teacher students' understanding of geometric optics concepts. The research sample consisted of 40 prospective physics teacher students, selected using the total sampling technique, with 20 students in each experimental and control group. The experimental group received learning with concreteness fading and multi-representation approaches, while the control group received conventional learning. The research instrument was a concept understanding test in the form of descriptions, which was tested for content validity and reliability using the inter-rater reliability method, yielding a Cohen's Kappa coefficient of 0.68. Data analysis techniques included independent samples t-test, ANCOVA, and N-gain. The results of the independent samples t-test showed that, in the posttest, there was a significant difference (p = 0.008 < 0.05) with the average score of the experimental group (M = 21.55) being higher than that of the control group (M = 19.10). ANCOVA test results showed that learning with concreteness fading-multi representation significantly affected students' concept understanding after controlling for the pretest score (p = 0.009 < 0.05). Additionally, the N-Gain test results indicated increased concept understanding in the experimental class (0.75, high category) and the control class (0.56, medium category). Initially, many students struggled and relied solely on one form of representation to explain geometric optics problems. However, after learning, they began to utilize various interrelated representations, including diagrams, texts, and mathematical equations. The findings in this study confirm that learning with concreteness fading and multi-representation approaches is effective in improving understanding of geometric optics concepts.      Keywords: conceptual understanding, concreteness fading, geometric optics, multi-representation, physics learning.