Abstract: Addressing the challenge of designing coal roadway support under deeply buried and complex geological conditions, this study integrates roadway support, computer development, and numerical simulation into a cohesive cross-fusion technological research approach. By compiling literature, conducting questionnaires, and undertaking field investigations, a comprehensive coal roadway support knowledge base was established. This base aids in providing users with essential rock mechanics and support scheme parameters through data retrieval and a rule-based reasoning mechanism. A versatile numerical simulation script, adaptable to varied geological conditions, diverse roadway types, and multiple support schemes, was developed utilizing FLAC^3D's built-in language for secondary development. This facilitated the creation of a predictive model for assessing the efficacy of coal roadway support systems. Advanced computer programming was employed to develop a predictive system for coal roadway support outcomes, enabling the visualization and intelligent forecasting of support effects. The system's efficacy was demonstrated through application to a classic case, with its predictive accuracy evaluated using the coefficient of determination (R²) and the Mean Absolute Percentage Error (MAPE), both yielding satisfactory results within acceptable ranges-R² values above 0.8 and MAPE values below 20%. Further validation in an engineering application, the 818 external air roadway at Xinhu Coal Mine, confirmed the system's high predictive accuracy of over 84%, proving its capability to provide accurate guidance for onsite construction operations.