The goal of this study was to determine the impact of Geometry Play Media on early childhood. The experimental group and the control group are used in this investigation. This is a quantitative research strategy using an Experimental Design and a Nonequivalent Control Group Design. The population in this study was 30 children of TK ABA 31 Semarang; the samples in class A were 15 children, and the samples in class B were 15 children. This research will use a saturated sampling technique. Saturated sampling is done if there are too few members of the population. Therefore, all members of the population are used as research samples. Techniques for gathering data include observation, documentation, and unstructured tests. The normality test, homogeneity test, and hypothesis test were employed in this study’s data analysis methodologies. According to the independent sample t-test calculation, the significance value of the control class pre-test is 0.001 < 0.05, and the value of the experimental class pre-test is 0.001 < 0.05. The significance value of the control class post-test is 0.000 < 0.05, and the value of the experimental class post-test is 0.000 < 0.05. As a result, the conclusion of the research conducted by researchers using Ho is rejected, whereas Ha is approved, indicating that the geometry media has a substantial influence on the numeracy skills of TK ABA 31 students in Semarang.