Abstract: To tackle the “black box” operational challenges in silicon iron alloy smelting process and the hard to improve the qualification rate of high-end grades because traditional quality control relies on manual experience, which lacks quantitative analysis methods, a full-process data governance and precise quality composition back-tracking solution is proposed. Firstly, a full-process data collection system for smelting, based on the “edge-cloud” collaborative architecture, is constructed. This system solves the spatiotemporal alignment and fusion difficulties in high-temperature, strongly coupled environments. Secondly, according to the principles of material balance and statistical process control, a mathematical back-tracking model for the “raw material-electrical-quality” is established, enabling the transformation of vague experience-based ironmaking into precise data mapping. Finally, a system with four core modules—data management, quality analysis, back-tracking query, and system management—is designed and developed. Engineering applications indicate that the system can effectively perform transparent storage and visual monitoring of full-process smelting data. With the precise back-tracking model, the quality anomaly attribution cycle is reduced from a few days to real-time, which makes product quality stability and batch consistency significantly improve. This research offers an effective technical approach for the digital transformation of silicon-based alloy smelting.