This study utilized Gaussian Graphical Models (GGM) and the NodeIdentifyR algorithm to investigate key forms of school adjustment problems among primary and middle school students. The GGM analysis revealed homework anxiety as a central form in both educational stages, with primary school students exhibiting behavioral problem and middle school students showing emotional problem. The NodeIdentifyR algorithm identified critical intervention nodes, revealing that unaddressed fighting behavior problems significantly exacerbates school adjustment problems in both stages. Targeted interventions, such as improving homework completion in primary school and reducing homework anxiety in middle school, were suggested as effective strategies. This study demonstrated an application of data-driven methods in tackling school adjustment problems. This approach aligns with the growing trend of integrating data science techniques in educational settings, offering a promising direction for enhancing student school adjustment.