<p>The accuracy of numerical simulation for steel ingot solidification significantly depends on the precise characterization of the interfacial heat transfer coefficient (IHTC). In this work, for three steel ingots with different weights of 2.8-ton, 19.0-ton, and 40.0-ton, the surface temperature of the ingot mold was measured in situ by an infrared thermal imager. Combined with the inverse calculation model in ProCAST software, the dynamic IHTC was solved, and a mathematical model of multi-ingot weight-dependent IHTC was constructed. The results show that the simulated surface temperature of the steel ingot using the IHTC back-calculated from measured temperatures is in high agreement with the experimental data. The IHTC exhibits a three-stage evolution pattern during solidification, it rapidly decreases from 3300 to 4000 W/(m²·°C) to approximately 500 W/(m²·°C) in the initial stage due to air gap formation, slowly declines to around 180 W/(m²·°C) in the middle stage with air gap expansion, and stabilizes at about 100 W/(m²·°C) in the late stage as the air gap becomes steady. Compared with the traditional fixed-value IHTC (2000 W/(m²·°C)), the back-calculated IHTC prolongs the total solidification time and reduces the temperature gradient of the 2.8-ton steel ingot. The width of the columnar crystal zone is shortened from 120 mm to 109 mm, and the central shrinkage volume is reduced by 31 pct. The IHTC model established through regression analysis shows that IHTC is linearly related to the ingot weight. Verified by an 8.5t steel ingot, the calculated values of the model are in good agreement with the measured data in the literature. A key boundary condition model for the precise simulation and process optimization of the steel ingot solidification process is provided by this study.</p>