With the continuous expansion of marine resource development， offshore platforms are evolving toward modularization and large-scale construction， imposing higher requirements on the relative motion control between the platform and the semi-submersible vessel during float-over operations. Traditional operation methods， which rely on manual experience and visual observation， struggle to meet the demands of refined operations. Based on the DGPS/INS integrated navigation system， this study develops an intelligent technical solution that integrates high-precision monitoring and motion prediction. Through centimeter-level positioning， visual modeling， and a multi-model optimized Kalman filtering algorithm， the system enables real-time monitoring of large offshore platform motion states and multi-step prediction of future motion trends. The system was successfully applied in the “Nan Hai Ba Hao” offshore platform float-over project. Field data demonstrate that the positioning accuracy reaches the centimeter level， and the motion prediction maintains high precision across different operational phases. The proposed solution provides accurate and real-time guidance and reliable data support for float-over operations， facilitating the transformation of operational decision-making from “experience-driven” to “data-driven” approach， thereby significantly enhancing operational safety and command efficiency.