This study explores the application of complex networks and computational science of science paradigms in the field of Science of Science, focusing on uncovering universal laws behind scientific activities while addressing the challenges of avoiding the “numerical game.” In the era of big data, large-scale data has provided new perspectives for science studies, where complex networks help reveal hidden patterns in scientific research by analyzing collaboration networks and knowledge dissemination pathways. Computational science of science, powered by advanced algorithms and models, has further enabled intelligent analysis in science studies. However, this data-driven approach also poses potential risks of falling into the “numerical game,” where over-reliance on quantitative analysis may overlook the complexities and contextual factors inherent in scientific activities. This study not only examines the strengths of complex networks and computational science of science in related studies but also delves into how to balance data-driven methods with domain knowledge to avoid the limitations of excessive data dependence. Ultimately, this paper aims to offer new perspectives for the future development of science studies and calls for more cautious and integrative approaches to address the challenges of modern science.