Abstract: In order to solve the problem that the outlet temperature of the strip steel in the seven stand of the finishing rolling area of one hot rolling line of a steel mill is affected by dynamic heat transfer interference, unsteady working conditions and measurement blind spots, which leads to the insufficient prediction accuracy of the traditional model, this paper combines a multiphysics coupled temperature prediction method. By constructing sub-models such as air-cooled radiation, water-cooled heat exchange, plastic deformation temperature rise, roll frictional heat effect and contact heat transfer, a temperature field calculation system considering the thermal-mechanical coupling effect is established. The inter-rack outlet temperature prediction software with integrated heat transfer coefficient correction algorithm was developed to realize the dynamic solution of the outlet temperature of each frame during the rolling process. Taking 650CL high-strength steel grade as the object, the temperature characteristic curve of specific heat capacity and thermal conductivity of steel grade was calculated by knot weighting method, and the experimental verification showed that the maximum deviation between the predicted value of the model and the infrared detection result was ≤ 10 ℃, and the relative error was < 1.5%. This technology effectively solves the problem of excessive error in the unsteady rolling stage of the traditional mechanism model, and provides reliable theoretical support for optimizing the rolling strategy and improving the dimensional accuracy and mechanical properties of high value-added products.