The accurate acquisition of the physical properties of strongly extinction materials limits their application development. In order to solve the problem of large errors in the high temperature measurement of these materials, this paper constructs an inversion model for the radiative and conductive coupled heat transfer through wavelet decomposition coefficients, and compares it with traditional methods. The radiative and conductive coupled heat transfer model is solved by finite volume method and Monte Carlo method. Daubechies nine wavelet is used for signal wavelet analysis. It has been discovered that the wavelet decomposition parameters of the temperature field are mainly concentrated in the 4–8 wavelet frequency band of wavelet, it can separate from the random error that is mainly concentrated in the 1–3 frequency band of wavelet. Most of physical properties like the heating power, convective heat transfer coefficient, material thermal conductivity, extinction coefficient has a great impact on the waveform and amplitude of wavelet detail coefficient and approximation coefficient, and the change trend of approximation coefficient and detail coefficient amplitude is consistent with the trend of temperature field. However, the effect of scattering albedo on wavelet parameters is tiny. Through inversion, it is found that the inversion model based on wavelet decomposition parameters of 5–7 frequency band has higher accuracy when random error exists.