This study investigates the effectiveness of business information management systems (BIMS) in cross-border real estate project management, using machine learning models for performance prediction. Data from 250 questionnaires and 15 interviews were analyzed using multiple regression, classification, and clustering methods. Results show that BIMS significantly improves project management efficiency and user satisfaction, reducing project completion time and cost overruns. Commercial real estate projects had the highest average investment ($70 million), while mixed-use projects had the longest average completion time (25 months). Residential projects achieved the highest management efficiency score (8.0). The regression model had an R² of 0.68, classification accuracy of 85%, and clustering identified three project types: high-efficiency, risk-concentrated, and resource-intensive. BIMS enhances decision-making through real-time data and risk identification. However, the study has limitations, including a small sample size and limited data sources. Future research should expand the sample and data sources to improve generalizability. The findings highlight the value of BIMS and machine learning in cross-border real estate management, offering practical tools for improving efficiency and risk management.This study investigates the effectiveness of business information management systems (BIMS) in cross-border real estate project management, using machine learning models for performance prediction. Data from 250 questionnaires and 15 interviews were analyzed using multiple regression, classification, and clustering methods. Results show that BIMS significantly improves project management efficiency and user satisfaction, reducing project completion time and cost overruns. Commercial real estate projects had the highest average investment ($70 million), while mixed-use projects had the longest average completion time (25 months). Residential projects achieved the highest management efficiency score (8.0). The regression model had an R² of 0.68, classification accuracy of 85%, and clustering identified three project types: high-efficiency, risk-concentrated, and resource-intensive. BIMS enhances decision-making through real-time data and risk identification. However, the study has limitations, including a small sample size and limited data sources. Future research should expand the sample and data sources to improve generalizability. The findings highlight the value of BIMS and machine learning in cross-border real estate management, offering practical tools for improving efficiency and risk management.