<p>The heating process of steel slabs is a critical stage in the production of medium and thick plates, significantly impacting production costs and energy efficiency. Existing research faces two major scientific challenges in gas consumption accounting and prediction for medium-thick steel slabs: (1) due to the dynamic movement of slabs, high-temperature environments, and structural constraints, it is difficult to directly measure slab temperatures and accurately quantify heat loss during the reheating process; and (2) traditional methods lack systematic consideration of batch integrity and the heat absorption ratio during the heating process, leading to low accuracy in energy consumption accounting and prediction. This paper proposes an innovative gas consumption accounting and prediction method based on batch integrity and heat absorption ratio, integrating metallurgical mechanisms and industrial big data. In the accounting phase, a 2D heat conduction finite difference method is used to calculate the heat absorption of slabs, and the batch energy consumption is allocated proportionally based on heat absorption ratios. In the prediction phase, XGBoost models are employed to predict batch gas consumption, with results distributed to individual slabs using the heat absorption ratio.</p>