This study, based on static data from over a thousand fracturing wells, employs data governance, data mining, and machine learning regression uncover princi­pal controlling factors for production in the fracturing context. Utilizing multiple evaluation methods, the entropy weight method comprehensively scores and ranks the principal controlling factors. A machine learning production prediction model is established for validation. Results show that DBSCAN achieves better accuracy in identifying field anomaly data. For missing data, it is recommended to use tree models or neural networks instead of imputation or constant filling, as incorrect imputation significantly degrades model performance. The entropy weight meth­od effectively integrates various correlation analysis results, providing a better connection with production compared to other approaches. This research utilizes large-scale field data to extract key parameters affecting production, supporting the establishment of high precision prediction models and the optimization of pa­rameters for unconventional reservoir production forecasts.