<p>As computational thinking (CT) becomes increasingly integrated into K–12 education worldwide, there is a growing need to understand the cognitive foundations that support its development. Spatial abilities have been recognized as an important factor related to CT. However, prior studies have often used single-task measures such as mental rotation to represent spatial skills, which oversimplifies their multidimensional nature. In addition, research on early elementary school children remains limited, making it difficult to understand how spatial abilities influence CT during this critical stage of development. This study addresses these gaps by examining the relationship between CT and four subdimensions of spatial ability (intrinsic static, intrinsic dynamic, extrinsic static, and extrinsic dynamic). A total of 237 students aged 6 to 10 completed a battery of tasks for assessing CT and different dimensions of spatial ability. Multiple linear regression analyses were conducted with CT performance as the dependent variable and the four spatial subdimensions as the independent variable, while age, gender, and coding experience were included as control variables to ensure the robustness of the results. The results showed that all four spatial subdimensions significantly predicted CT performance, even after controlling for the covariates. These findings contribute to a deeper understanding of the cognitive basis of CT by demonstrating the role of diverse spatial skills in its development. The study highlights the importance of incorporating varied types of spatial training in early education and provides empirical support for integrating spatial learning into STEM curricula.</p>