PROBLEM TO BE SOLVED: To provide a thermal conductivity estimation method capable of easily estimating thermal conductivity in conducting various heat transfer analyzes in a manufacturing process of a semiconductor crystal product. SOLUTION: A thermal conductivity estimating method comprises a step of heating a part of a measurement sample under a predetermined heating condition to measure a temperature distribution on a surface of the measurement sample in a steady state, and a sample model having the same shape as the measurement sample. A step of performing heat transfer simulation for a plurality of combinations of the provisional thermal conductivity and the heating condition, and calculating the temperature distribution on the sample model surface for each combination, and the plurality of combinations and the temperatures obtained from the plurality of combinations. Using the machine learning method to create a regression model that uses the distribution calculation results as training data, the input as the temperature distribution on the measurement sample surface, and the output as the thermal conductivity of the measurement sample, and the measurement of the temperature distribution on the measurement sample surface. Inputting the result to the regression model and estimating the thermal conductivity of the measurement sample. [Selection diagram] Figure 1
展开▼