This study aims to determine the effectiveness of GeoGebra-assisted learning on students' mathematical representation through a meta-analysis study and possible moderator variables to moderate it. Experimental studies were obtained from Google Scholar and ERIC databases and then included in this meta-analysis study were 23 studies that met the inclusion criteria. The effect sizes were combined according to the Random Effects Model (REM). Based on the calculation using CMA v3 and R Studio (Meta Package), the overall effect size was calculated as 1.144,95% [0.890, 1.397]. Furthermore, the application of GeoGebra-assisted learning to improve students' mathematical representation ability is also influenced by students' demographics (Sumatra Island) (Q = 7.874 > X[superscript 2], p < 0.05). However, the application of GeoGebra-assisted learning to improve students' mathematical representation skills was not influenced by variables such as experimental design (Q = 0.306 < X[superscript 2], p > 0.05), independent variables (Q = 1.438 < X[superscript 2], p > 0.05), publication type (Q = 0.674 < X[superscript 2], p > 0.05), education level (Q = 1.619 < X[superscript 2], p > 0.05), subject matter (Q = 3.330 < X[superscript 2], p > 0.05), meeting duration (Q = 0.25 < X[superscript 2], p > 0.05), and sample size (Q = 0.53 < X[superscript 2], p > 0.05). Based on these results, it can be said that GeoGebra-assisted learning has a strong positive effect on students' mathematical representation when compared to learning without GeoGebra (conventional).