The hearth thermal state is an important index to reflect the operation state of blast furnace hearth, which has guiding significance for the operation of blast furnace (BF) in high yield, energy saving and emission reduction. The iron-making is carried out in a closed container with high temperature and pressure. It is a complex industrial process with multivariable, distributed, strong coupling and time-varying condition. In addition, the production conditions of the blast furnace often fluctuate, resulting in the same state does not necessarily mean that the temperature of the hearth is same. The comprehensive evaluation of hearth thermal state by the blast furnace operator is inevitably affected by personal subjective factors, it also has problems of uncertainty and fuzziness. Based on the furnace temperature prediction, according to the actual production situation and expert experience, the important parameters which represented the thermal state of the blast furnace hearth were analyzed and extracted, and the comprehensive evaluation index system of the hearth thermal state was established in this work. Then, a two-stage fuzzy comprehensive evaluation model of blast furnace hearth thermal state (too low, low, suitable, high, too high) by using statistical method was built to determine the weights of evaluation indexes and the membership degree function. Finally, the comprehensive evaluation of hearth thermal state to 2500 m3blast furnace was applied to evaluate the thermal state of hearth comprehensively. At the same time, the evaluation was used to compare with the actual operation situation. From the comparison results, there were 794 groups matched perfectly (d=0) in the evaluation results of the model, accounting for 74.21%. There were 245 groups matched reliably (d=±1), accounting for 22.90%, and the matching rate reached 97.11%. The evaluation model provided an accurate and reliable basis for realizing the stable thermal state of the blast furnace hearth and producing energy saving and emission reduction.